Current audio-driven facial animation methods achieve impressive results for short videos but suffer from error accumulation and identity drift when extended to longer durations. Existing methods attempt to mitigate this through external spatial control, increasing long-term consistency but compromising the naturalness of motion. We propose KeyFace, a novel two-stage diffusion-based framework, to address these issues. In the first stage, keyframes are generated at a low frame rate, conditioned on audio input and an identity frame, to capture essential facial expressions and movements over extended periods of time. In the second stage, an interpolation model fills in the gaps between keyframes, ensuring smooth transitions and temporal coherence. To further enhance realism, we incorporate continuous emotion representations and handle a wide range of non-speech vocalizations (NSVs), such as laughter and sighs. We also introduce two evaluation metrics for assessing lip synchronization and NSV generation. Experimental results show that KeyFace outperforms state-of-the-art methods in generating natural, coherent facial animations over extended durations, successfully encompassing NSVs and continuous emotions. See the project page for visualizations and code.

Multimodal large language models (MLLMs) have recently become a focal point of research due to their formidable multimodal understanding capabilities. For example, in the audio and speech domains, an LLM can be equipped with (automatic) speech recognition (ASR) abilities by just concatenating the audio tokens, computed with an audio encoder, and the text tokens to achieve state-of-the-art results. On the contrary, tasks like visual and audio-visual speech recognition (VSR/AVSR), which also exploit noise-invariant lip movement information, have received little or no attention. To bridge this gap, we propose Llama-AVSR, a new MLLM with strong audio-visual speech recognition capabilities. It leverages pre-trained audio and video encoders to produce modality-specific tokens which, together with the text tokens, are processed by a pre-trained LLM (e.g., Llama3.1-8B) to yield the resulting response in an auto-regressive fashion. Llama-AVSR requires a small number of trainable parameters as only modality-specific projectors and LoRA modules are trained whereas the multi-modal encoders and LLM are kept frozen. We evaluate our proposed approach on LRS3, the largest public AVSR benchmark, and we achieve new state-of-the-art results for the tasks of ASR and AVSR with a WER of 0.79% and 0.77%, respectively. To bolster our results, we investigate the key factors that underpin the effectiveness of Llama-AVSR: the choice of the pre-trained encoders and LLM, the efficient integration of LoRA modules, and the optimal performance-efficiency trade-off obtained via modality-aware compression rates.

This paper investigates the under-explored area of low-rank weight training for large-scale Conformer-based speech recognition models from scratch. Our study demonstrates the viability of this training paradigm for such models, yielding several notable findings. Firstly, we discover that applying a low-rank structure exclusively to the attention modules can unexpectedly enhance performance, even with a significant rank reduction of 12%. In contrast, feed-forward layers present greater challenges, as they begin to exhibit performance degradation with a moderate 50% rank reduction. Furthermore, we find that both initialization and layer-wise rank assignment play critical roles in successful low-rank training. Specifically, employing SVD initialization and linear layer-wise rank mapping significantly boosts the efficacy of low-rank weight training. Building on these insights, we introduce the Low-Rank Speech Model from Scratch (LR-SMS), an approach that achieves performance parity with full-rank training while delivering substantial reductions in parameters count (by at least 2×), and training time speedups (by 1.3× for ASR and 1.15× for AVSR).

In this paper, we investigate a novel approach for Target Speech Extraction (TSE), which relies solely on textual context to extract the target speech. We refer to this task as Contextual Speech Extraction (CSE). Unlike traditional TSE methods that rely on pre-recorded enrollment utterances, video of the target speaker’s face, spatial information, or other explicit cues to identify the target stream, our proposed method requires only a few turns of previous dialogue (or monologue) history. This approach is naturally feasible in mobile messaging environments where voice recordings are typically preceded by textual dialogue that can be leveraged implicitly. We present three CSE models and analyze their performances on three datasets. Through our experiments, we demonstrate that even when the model relies purely on dialogue history, it can achieve over 90% accuracy in identifying the correct target stream with only two previous dialogue turns. Furthermore, we show that by leveraging both textual context and enrollment utterances as cues during training, we further enhance our model’s flexibility and effectiveness, allowing us to use either cue during inference, or combine both for improved performance.

Large language models (LLMs) have recently shown strong potential in audio-visual speech recognition (AVSR), but their high computational demands and sensitivity to token granularity limit their practicality in resource-constrained settings. Token compression methods can reduce inference cost, but they require fixing a compression rate in advance and produce a single fixed-length output, offering no flexibility to balance information density and efficiency at inference time. Matryoshka representation learning (MRL) addresses this by enabling a single model to operate across multiple token granularities, allowing compression rates to be adjusted dynamically. However, current MRL-based methods treat each scale independently during training, limiting cross-scale generalization, robustness at high compression, and interpretability. To overcome these limitations, we propose MoME (Mixture of Matryoshka Experts), a novel framework that integrates sparse Mixture-of-Experts (MoE) into MRL-based LLMs for AVSR. MoME augments a frozen LLM with top-k routed and shared experts, allowing dynamic capacity allocation across scales and modalities. A shared router promotes consistent expert activation across granularities, enabling compressed sequences to benefit from representations learned at lower compression. Experiments on LRS2 and LRS3 demonstrate that MoME achieves state-of-the-art performance across AVSR, ASR, and VSR tasks, while requiring significantly fewer parameters and maintaining robustness under noise. MoME unifies the adaptability of MRL with the efficiency of MoE, offering a scalable and interpretable solution for resource-aware speech recognition.

Image-based age estimation aims to predict a person's age from facial images. It is used in a variety of real-world applications. Although end-to-end deep models have achieved impressive results for age estimation on benchmark datasets, their performance in-the-wild still leaves much room for improvement due to the challenges caused by large variations in head pose, facial expressions, and occlusions. To address this issue, we propose a simple yet effective method to explicitly incorporate facial semantics into age estimation, so that the model would learn to correctly focus on the most informative facial components from unaligned facial images regardless of head pose and non-rigid deformation. To this end, we design a face parsing-based network to learn semantic information at different scales and a novel face parsing attention module to leverage these semantic features for age estimation. To evaluate our method on in-the-wild data, we also introduce a new challenging large-scale benchmark called IMDB-Clean. This dataset is created by semi-automatically cleaning the noisy IMDB-WIKI dataset using a constrained clustering method. Through comprehensive experiment on IMDB-Clean and other benchmark datasets, under both intra-dataset and cross-dataset evaluation protocols, we show that our method consistently outperforms all existing age estimation methods and achieves a new state-of-the-art performance. To the best of our knowledge, our work presents the first attempt of leveraging face parsing attention to achieve semantic-aware age estimation, which may be inspiring to other high level facial analysis tasks.

Head avatars animated by visual signals have gained popularity, particularly in cross-driving synthesis where the driver differs from the animated character, a challenging but highly practical approach. The recently presented MegaPortraits model has demonstrated state-of-the-art results in this domain. We conduct a deep examination and evaluation of this model, with a particular focus on its latent space for facial expression descriptors, and uncover several limitations with its ability to express intense face motions. To address these limitations, we propose substantial changes in both training pipeline and model architecture, to introduce our EMOPortraits model, where we: Enhance the model's capability to faithfully support intense, asymmetric face expressions, setting a new state-of-the-art result in the emotion transfer task, surpassing previous methods in both metrics and quality. Incorporate speech-driven mode to our model, achieving top-tier performance in audio-driven facial animation, making it possible to drive source identity through diverse modalities, including visual signal, audio, or a blend of both. Furthermore, we propose a novel multi-view video dataset featuring a wide range of intense and asymmetric facial expressions, filling the gap with absence of such data in existing datasets.

Self-supervision has recently shown great promise for learning visual and auditory speech representations from unlabelled data. In this work, we propose BRAVEn, an extension to the recent RAVEn method, which learns speech representations entirely from raw audio-visual data. Our modifications to RAVEn enable BRAVEn to achieve state-of-the-art results among self-supervised methods in various settings. Moreover, we observe favourable scaling behaviour by increasing the amount of unlabelled data well beyond other self-supervised works. In particular, we achieve 20.0% / 1.7% word error rate for VSR / ASR on the LRS3 test set, with only 30 hours of labelled data and no external ASR models. Our results suggest that readily available unlabelled audiovisual data can largely replace costly transcribed data. Code at https://github.com/ahaliassos/raven.

The recent success of Automatic Speech Recognition (ASR) is largely attributed to the ever-growing amount of training data. However, this trend has made model training prohibitively costly and imposed computational demands. While data pruning has been proposed to mitigate this issue by identifying a small subset of relevant data, its application in ASR has been barely explored, and existing works often entail significant overhead to achieve meaningful results. To fill this gap, this paper presents the first investigation of dynamic data pruning for ASR, finding that we can reach the full-data performance by dynamically selecting 70% of data. Furthermore, we introduce Dynamic Data Pruning for ASR (DDP-ASR), which offers several fine-grained pruning granularities specifically tailored for speech-related datasets, going beyond the conventional pruning of entire time sequences. Our intensive experiments show that DDP-ASR can save up to 1.6x training time with negligible performance loss.

Research in auditory, visual, and audiovisual speech recognition (ASR, VSR, and AVSR, respectively) has traditionally been conducted independently. Even recent self-supervised studies addressing two or all three tasks simultaneously tend to yield separate models, leading to disjoint inference pipelines with increased memory requirements and redundancies. This paper proposes unified training strategies for these systems. We demonstrate that training a single model for all three tasks enhances VSR and AVSR performance, overcoming typical optimisation challenges when training from scratch. Moreover, we introduce a greedy pseudo-labelling approach to more effectively leverage unlabelled samples, addressing shortcomings in related self-supervised methods. Finally, we develop a self-supervised pre-training method within our framework, proving its effectiveness alongside our semi-supervised approach. Despite using a single model for all tasks, our unified approach achieves state-of-the-art performance on LRS3 for ASR, VSR, and AVSR compared to recent methods. Code will be made publicly available.

Talking face generation has historically struggled to produce head movements and natural facial expressions without guidance from additional reference videos. Recent developments in diffusion-based generative models allow for more realistic and stable data synthesis and their performance on image and video generation has surpassed that of other generative models. In this work, we present an autoregressive diffusion model that requires only one identity image and audio sequence to generate a video of a realistic talking head. Our solution is capable of hallucinating head movements, facial expressions, such as blinks, and preserving a given background. We evaluate our model on two different datasets, achieving state-of-the-art results in expressiveness and smoothness on both of them.1

Video-to-speech synthesis is the task of reconstructing the speech signal from a silent video of a speaker. Previous approaches train on data from almost exclusively audio-visual datasets, i.e., every audio sample has a corresponding video sample. This precludes the use of abundant audio-only datasets which may not have a corresponding visual modality such as audiobooks, radio podcasts, and speech recognition datasets. In this paper we propose to train encoder-decoder models on more than 3,500 hours of audio data at 24kHz, and then use the pretrained decoders to initialize the audio decoders for the video-tospeech synthesis task. The pre-training step uses audio samples only and does not require labels or corresponding samples from other modalities (visual, text). We demonstrate that this improves the reconstructed speech and that it is an unexplored way to improve the quality of the generator in a cross-modal task while only requiring samples from one of the modalities. We conduct experiments using both raw audio and mel spectrograms as target outputs and benchmark our models with existing work.

Recently reported state-of-the-art results in visual speech recognition (VSR) often rely on increasingly large amounts of video data, while the publicly available tran-scribed video datasets are limited in size. In this paper, for the first time, we study the potential of leveraging synthetic visual data for VSR. Our method, termed Synth VSR, sub-stantially improves the performance of VSR systems with synthetic lip movements. The key idea behind Synth VSR is to leverage a speech-driven lip animation model that gen-erates lip movements conditioned on the input speech. The speech-driven lip animation model is trained on an unla-beled audio-visual dataset and could be further optimized towards a pre-trained VSR model when labeled videos are available. As plenty of transcribed acoustic data and face images are available, we are able to generate large-scale synthetic data using the proposed lip animation model for semi-supervised VSR training. We evaluate the performance of our approach on the largest public VSR bench-mark - Lip Reading Sentences 3 (LRS3). Synth VSR achieves a WER of 43.3% with only 30 hours of real labeled data, outperforming off-the-shelf approaches using thousands of hours of video. The WER is further reduced to 27.9% when using all 438 hours of labeled data from LRS3, which is on par with the state-of-the-art self-supervised AV-HuBERT method. Furthermore, when combined with large-scale pseudo-labeled audio-visual data SynthVSR yields a new state-of-the-art VSR WER of 16.9% using publicly available data only, surpassing the recent state-of-the-art approaches trained with 29 times more non-public machine-transcribed video data (90,000 hours). Finally, we perform extensive ablation studies to understand the effect of each component in our proposed method.

This work focuses on the apparent emotional reaction recognition (AERR) from the video-only input, conducted in a self-supervised fashion. The network is first pre-trained on different self-supervised pretext tasks and later fine-tuned on the downstream target task. Self-supervised learning facilitates the use of pre-trained architectures and larger datasets that might be deemed unfit for the target task and yet might be useful to learn informative representations and hence provide useful initializations for further fine-tuning on smaller more suitable data. Our presented contribution is two-fold: (1) an analysis of different state-of-the-art (SOTA) pretext tasks for the video-only apparent emotional reaction recognition architecture, and (2) an analysis of various combinations of the regression and classification losses that are likely to improve the performance further. Together these two contributions result in the current state-of-the-art performance for the video-only spontaneous apparent emotional reaction recognition with continuous annotations.

Audio-visual speech recognition has received a lot of attention due to its robustness against acoustic noise. Recently, the performance of automatic, visual, and audio-visual speech recognition (ASR, VSR, and AV-ASR, respectively) has been substantially improved, mainly due to the use of larger models and training sets. However, accurate labelling of datasets is time-consuming and expensive. Hence, in this work, we investigate the use of automatically-generated transcriptions of unlabelled datasets to increase the training set size. For this purpose, we use publicly-available pre-trained ASR models to automatically transcribe unlabelled datasets such as AVSpeech and Vox-Celeb2. Then, we train ASR, VSR and AV-ASR models on the augmented training set, which consists of the LRS2 and LRS3 datasets as well as the additional automatically-transcribed data. We demonstrate that increasing the size of the training set, a recent trend in the literature, leads to reduced WER despite using noisy transcriptions. The proposed model achieves new state-of-the-art performance on AV-ASR on LRS2 and LRS3. In particular, it achieves a WER of 0.9 % on LRS3, a relative improvement of 30 % over the current state-of-the–art approach, and outperforms methods that have been trained on non-publicly available datasets with 26 times more training data.

Audio-visual speech enhancement aims to extract clean speech from a noisy environment by leveraging not only the audio itself but also the target speaker’s lip movements. This approach has been shown to yield improvements over audio-only speech enhancement, particularly for the removal of interfering speech. Despite recent advances in speech synthesis, most audio-visual approaches continue to use spectral mapping/masking to reproduce the clean audio, often resulting in visual backbones added to existing speech enhancement architectures. In this work, we propose LA-VocE, a new two-stage approach that predicts mel-spectrograms from noisy audio-visual speech via a transformer-based architecture, and then converts them into waveform audio using a neural vocoder (HiFi-GAN). We train and evaluate our framework on thousands of speakers and 11+ different languages, and study our model’s ability to adapt to different levels of background noise and speech interference. Our experiments show that LA-VocE outperforms existing methods according to multiple metrics, particularly under very noisy scenarios.

Cross-lingual self-supervised learning has been a growing research topic in the last few years. However, current works only explored the use of audio signals to create representations. In this work, we study cross-lingual self-supervised visual representation learning. We use the recently-proposed Raw Audio-Visual Speech Encoders (RAVEn) framework to pre-train an audio-visual model with unlabelled multilingual data, and then fine-tune the visual model on labelled transcriptions. Our experiments show that: (1) multi-lingual models with more data outperform monolingual ones, but, when keeping the amount of data fixed, monolingual models tend to reach better performance; (2) multi-lingual outperforms English-only pre-training; (3) using languages which are more similar yields better results; and (4) fine-tuning on unseen languages is competitive to using the target language in the pre-training set. We hope our study inspires future research on non-English-only speech representation learning.

Due to its importance in facial behaviour analysis, facial action unit (AU) detection has attracted increasing attention from the research community. Leveraging the online knowledge distillation framework, we propose the "FAN-Trans" method for AU detection. Our model consists of a hybrid network of convolution and transformer blocks to learn per-AU features and to model AU co-occurrences. The model uses a pre-trained face alignment network as the feature extractor. After further transformation by a small learnable add-on convolutional subnet, the per-AU features are fed into transformer blocks to enhance their representation. As multiple AUs often appear together, we propose a learnable attention drop mechanism in the transformer block to learn the correlation between the features for different AUs. We also design a classifier that predicts AU presence by considering all AUs’ features, to explicitly capture label dependencies. Finally, we make the attempt of adapting online knowledge distillation in the training stage for this task, further improving the model’s performance. Experiments on the BP4D and DISFA datasets demonstrating the effectiveness of proposed method.

Human-computer interaction is becoming increasingly prevalent in daily life with the adoption of intelligent devices. These devices must be capable of interacting in diverse settings, such as environments with noise, music and differing illumination and occlusion conditions. They must also interact with a variety of end users across ages and backgrounds. Therefore, the machine learning community needs in-the-wild multi-modal datasets to develop models for face and speech analysis so that they can be applicable in most real world scenarios. However, most existing audio and audio-visual databases are captured in controlled conditions with few or no age and kinship labels. In this paper, we introduce the KAN-AV dataset which contains 98 h of audio-visual data from 970 identities across ages. Two thirds of the identities have kin relations in the dataset. The dataset is manually annotated with labels for kinship, age, and gender and is intended to drive future research in face and speech analysis.

This article presents a novel method for face clustering in videos using a video-centralised transformer. Previous works often employed contrastive learning to learn frame-level representation and used average pooling to aggregate the features along the temporal dimension. This approach may not fully capture the complicated video dynamics. In addition, despite the recent progress in video-based contrastive learning, few have attempted to learn a self-supervised clustering-friendly face representation that benefits the video face clustering task. To overcome these limitations, our method employs a transformer to directly learn video-level representations that can better reflect the temporally-varying property of faces in videos, while we also propose a video-centralised self-supervised framework to train the transformer model. We also investigate face clustering in egocentric videos, a fast-emerging field that has not been studied yet in works related to face clustering. To this end, we present and release the first large-scale egocentric video face clustering dataset named EasyCom-Clustering. We evaluate our proposed method on both the widely used Big Bang Theory (BBT) dataset and the new EasyCom-Clustering dataset. Results show the performance of our video-centralised transformer has surpassed all previous state-of-the-art methods on both benchmarks, exhibiting a self-attentive understanding of face videos.

Facial action unit (AU) detection plays an important role in performing facial behavioral analysis of raw video inputs. Overall, there are three key factors that contribute toward the optimal performance of AU detectors: 1) being able to capture local AU-centered features; 2) exploiting the fact that some AUs co-occur with others; and 3) utilizing appearance changes across frames. We briefly review current techniques addressing each factor and discuss the challenges they meet. Given that very few works consider how to effectively and efficiently merge them all into a single framework that can be trained in an end-to-end manner, we propose facial AU detection with face alignment(AUNet), a simple yet strong baseline for landmark-based AU detection. AUNet implements the abovementioned key factors by: 1) using the intermediate layers of a pretrained face alignment model to act as our AU features’ space; 2) optimized to satisfy a correlation constraint, derived from the AU labels, and 3) temporal constraint, derived from variations in the contents of consecutive frames in the input videos. The proposed model, with its three key components, remains simple in nature and aligns with the primary AU detection task. Experiments on several benchmarks show that it substantially improves the AU detector’s accuracy and achieves new state-of-the-art AU detection results on popular benchmarks: BP4D and DISFA. Code is available athttps://github.com/jingyang2017/AU-Net.

Micro-expressions (ME) are a special form of facial expressions which may occur when people try to hide their true feelings for some reasons. MEs are important clues to reveal people’s true feelings, but are difficult or impossible to be captured by ordinary persons with naked-eyes as they are very short and subtle. It is expected that robust computer vision methods can be developed to automatically analyze MEs which requires lots of ME data. The current ME datasets are insufficient, and mostly contain only one single form of 2D color videos. Researches on 4D data of ordinary facial expressions have prospered, but so far no 4D data is available in ME study. In the current study, we introduce the 4DME dataset: a new spontaneous ME dataset which includes 4D data along with three other video modalities. Both micro- and macro-expression clips are labeled out in 4DME, and 22 AU labels and five categories of emotion labels are annotated. Experiments are carried out using three 2D-based methods and one 4D-based method to provide baseline results. The results indicate that the 4D data can potentially benefit ME recognition. The 4DME dataset could be used for developing 4D-based approaches, or exploring fusion of multiple video sources (e.g., texture and depth) for the task of ME analysis in future. Besides, we also emphasize the importance of forming a clear and unified criteria of ME annotation for future ME data collection studies. Several key questions related with ME annotation are listed and discussed in depth, especially about the relationship between AUs and ME emotion categories. A preliminary AU-Emo mapping table is proposed with justified explanations and supportive experimental results. Several unsolved issues are also summarized for future work.

Self-supervised learning has attracted plenty of recent research interest. However, most works for self-supervision in speech are typically unimodal and there has been limited work that studies the interaction between audio and visual modalities for cross-modal self-supervision. This article (1) investigates visual self-supervision via face reconstruction to guide the learning of audio representations; (2) proposes an audio-only self-supervision approach for speech representation learning; (3) shows that a multi-task combination of the proposed visual and audio self-supervision is beneficial for learning richer features that are more robust in noisy conditions; (4) shows that self-supervised pretraining can outperform fully supervised training and is especially useful to prevent overfitting on smaller sized datasets. We evaluate our learned audio representations for discrete emotion recognition, continuous affect recognition and automatic speech recognition. We outperform existing self-supervised methods for all tested downstream tasks. Our results demonstrate the potential of visual self-supervision for audio feature learning and suggest that joint visual and audio self-supervision leads to more informative audio representations for speech and emotion recognition.

Video-to-speech is the process of reconstructing the audio speech from a video of a spoken utterance. Previous approaches to this task have relied on a two-step process where an intermediate representation is inferred from the video and is then decoded into waveform audio using a vocoder or a waveform reconstruction algorithm. In this work, we propose a new end-to-end video-to-speech model based on generative adversarial networks (GANs) which translates spoken video to waveform end-to-end without using any intermediate representation or separate waveform synthesis algorithm. Our model consists of an encoder-decoder architecture that receives raw video as input and generates speech, which is then fed to a waveform critic and a power critic. The use of an adversarial loss based on these two critics enables the direct synthesis of the raw audio waveform and ensures its realism. In addition, the use of our three comparative losses helps establish direct correspondence between the generated audio and the input video. We show that this model is able to reconstruct speech with remarkable realism for constrained datasets such as GRID, and that it is the first end-to-end model to produce intelligible speech for Lip Reading in the Wild (LRW), featuring hundreds of speakers recorded entirely "in the wild." We evaluate the generated samples in two different scenarios-seen and unseen speakers-using four objective metrics which measure the quality and intelligibility of artificial speech. We demonstrate that the proposed approach outperforms all previous works in most metrics on GRID and LRW.

One of the most pressing challenges for the detection of face-manipulated videos is generalising to forgery methods not seen during training while remaining effective under common corruptions such as compression. In this paper, we examine whether we can tackle this issue by harnessing videos of real talking faces, which contain rich information on natural facial appearance and behaviour and are readily available in large quantities online. Our method, termed RealForensics, consists of two stages. First, we exploit the natural correspondence between the visual and auditory modalities in real videos to learn, in a self-supervised cross-modal manner, temporally dense video representations that capture factors such as facial movements, expression, and identity. Second, we use these learned representations as targets to be predicted by our forgery detector along with the usual binary forgery classification task; this encourages it to base its real/fake decision on said factors. We show that our method achieves state-of-the-art performance on cross-manipulation generalisation and robustness experiments, and examine the factors that contribute to its per-formance. Our results suggest that leveraging natural and unlabelled videos is a promising direction for the development of more robust face forgery detectors.

Several training strategies and temporal models have been recently proposed for isolated word lip-reading in a series of independent works. However, the potential of combining the best strategies and investigating the impact of each of them has not been explored. In this paper, we systematically investigate the performance of state-of-the-art data augmentation approaches, temporal models and other training strategies, like self-distillation and using word boundaries indicators. Our results show that time masking is the most important augmentation followed by mixup and Densely-Connected Temporal Convolutional Networks (DC-TCN) are the best temporal model for lip-reading of isolated words. Using self-distillation and word boundaries is also beneficial but to a lesser extent. A combination of all the above methods results in a classification accuracy of 93.4%, which is an absolute improvement of 4.6% over the current state-of-the-art performance on the LRW dataset. The performance can be further improved to 94.1% by pre-training on additional datasets. An error analysis of the various training strategies reveals that the performance improves by increasing the classification accuracy of hard-to-recognise words.

Video-to-speech synthesis (also known as lip-to-speech) refers to the translation of silent lip movements into the corresponding audio. This task has received an increasing amount of attention due to its self-supervised nature (i.e., can be trained without manual labelling) combined with the ever-growing collection of audio-visual data available online. Despite these strong motivations, contemporary video-to-speech works focus mainly on small- to medium-sized corpora with substantial constraints in both vocabulary and setting. In this work, we introduce a scalable video-to-speech framework consisting of two components: a video-to-spectrogram predictor and a pre-trained neural vocoder, which converts the mel-frequency spectrograms into waveform audio. We achieve state-of-the art results for GRID and considerably outperform previous approaches on LRW. More importantly, by focusing on spectrogram prediction using a simple feedforward model, we can efficiently and effectively scale our method to very large and unconstrained datasets: To the best of our knowledge, we are the first to show intelligible results on the challenging LRS3 dataset.

Recent work in unsupervised learning has focused on efficient inference and learning in latent variables models. Training these models by maximizing the evidence (marginal likelihood) is typically intractable. Thus, a common approximation is to maximize the Evidence Lower BOund (ELBO) instead. Variational autoencoders (VAE) are a powerful and widely-used class of generative models that optimize the ELBO efficiently for large datasets. However, the VAE's default Gaussian choice for the prior imposes a strong constraint on its ability to represent the true posterior, thereby degrading overall performance. A Gaussian mixture model (GMM) would be a richer prior but cannot be handled efficiently within the VAE framework because of the intractability of the Kullback-Leibler divergence for GMMs. We deviate from the common VAE framework in favor of one with an analytical solution for Gaussian mixture prior. To perform efficient inference for GMM priors, we introduce a new constrained objective based on the Cauchy-Schwarz divergence, which can be computed analytically for GMMs. This new objective allows us to incorporate richer, multi-modal priors into the autoencoding framework. We provide empirical studies on a range of datasets and show that our objective improves upon variational auto-encoding models in density estimation, unsupervised clustering, semi-supervised learning, and face analysis.

Although current deep learning-based face forgery detectors achieve impressive performance in constrained scenarios, they are vulnerable to samples created by unseen manipulation methods. Some recent works show improvements in generalisation but rely on cues that are easily corrupted by common post-processing operations such as compression. In this paper, we propose LipForensics, a detection approach capable of both generalising to novel manipulations and withstanding various distortions. LipForensics targets high-level semantic irregularities in mouth movements, which are common in many generated videos. It consists in first pretraining a spatio-temporal network to perform visual speech recognition (lipreading), thus learning rich internal representations related to natural mouth motion. A temporal network is subsequently finetuned on fixed mouth embeddings of real and forged data in order to detect fake videos based on mouth movements without overfitting to low-level, manipulation-specific artefacts. Extensive experiments show that this simple approach significantly surpasses the state-of-the-art in terms of generalisation to unseen manipulations and robustness to perturbations, as well as shed light on the factors responsible for its performance.

Lipreading has witnessed a lot of progress due to the resurgence of neural networks. Recent works have placed emphasis on aspects such as improving performance by finding the optimal architecture or improving generalization. However, there is still a significant gap between the current methodologies and the requirements for an effective deployment of lipreading in practical scenarios. In this work, we propose a series of innovations that significantly bridge that gap: first, we raise the state-of-the-art performance by a wide margin on LRW and LRW-1000 to 88.5 % and 46.6 %, respectively using self-distillation. Secondly, we propose a series of architectural changes, including a novel Depthwise Separable Temporal Convolutional Network (DS-TCN) head, that slashes the computational cost to a fraction of the (already quite efficient) original model. Thirdly, we show that knowledge distillation is a very effective tool for recovering performance of the lightweight models. This results in a range of models with different accuracy-efficiency trade-offs. However, our most promising lightweight models are on par with the current state-of-the-art while showing a reduction of 8.2× and 3.9× in terms of computational cost and number of parameters, respectively, which we hope will enable the deployment of lipreading models in practical applications.

Adversarial attacks pose a threat to deep learning models. However, research on adversarial detection methods, especially in the multi-modal domain, is very limited. In this work, we propose an efficient and straightforward detection method based on the temporal correlation between audio and video streams. The main idea is that the correlation between audio and video in adversarial examples will be lower than benign examples due to added adversarial noise. We use the synchronisation confidence score as a proxy for audiovisual correlation and based on it we can detect adversarial attacks. To the best of our knowledge, this is the first work on detection of adversarial attacks on audiovisual speech recognition models. We apply recent adversarial attacks on two audiovisual speech recognition models trained on the GRID and LRW datasets. The experimental results demonstrate that the proposed approach is an effective way for detecting such attacks.

In this work, we present a hybrid CTC/Attention model based on a ResNet-18 and Convolution-augmented transformer (Conformer), that can be trained in an end-to-end manner. In particular, the audio and visual encoders learn to extract features directly from raw pixels and audio waveforms, respectively, which are then fed to conformers and then fusion takes place via a Multi-Layer Perceptron (MLP). The model learns to recognise characters using a combination of CTC and an attention mechanism. We show that end-to-end training, instead of using pre-computed visual features which is common in the literature, the use of a conformer, instead of a recurrent network, and the use of a transformer-based language model, significantly improve the performance of our model. We present results on the largest publicly available datasets for sentence-level speech recognition, Lip Reading Sentences 2 (LRS2) and Lip Reading Sentences 3 (LRS3), respectively. The results show that our proposed models raise the state-of-the-art performance by a large margin in audio-only, visual-only, and audio-visual experiments.

State-of-the-art audio classification systems often apply deep neural networks on hand-crafted features (e.g., spectrogram-based representations), instead of learning features directly from raw audio. Moreover, these audio networks have millions of unknown parameters need to be learned, which causes a great demand for computational resources and training data. In this paper, we aim to learn audio representations directly from raw audio, and at the same time mitigate its training burden by employing a light-weight architecture. In particular, we propose to learn separable filters, parametrized with only a few variables, namely center frequency and bandwidth, facilitating training and offering interpretability of learned representations. The generality of the proposed method is demonstrated by applying it onto two applications, namely 1) speaker identification and 2) acoustic event recognition. Experimental results indicate its effectiveness on these applications, especially when small amount of training data is available.

The large amount of audiovisual content being shared online today has drawn substantial attention to the prospect of audiovisual self-supervised learning.Recent works have focused on each of these modalities separately, while others have attempted to model both simultaneously in a cross-modal fashion.However, comparatively little attention has been given to leveraging one modality as a training objective to learn from the other.In this work, we propose Learning visual speech Representations from Audio via self-supervision (LiRA).Specifically, we train a ResNet+Conformer model to predict acoustic features from unlabelled visual speech.We find that this pre-trained model can be leveraged towards word-level and sentence-level lip-reading through feature extraction and fine-tuning experiments.We show that our approach significantly outperforms other self-supervised methods on the Lip Reading in the Wild (LRW) dataset and achieves state-of-the-art performance on Lip Reading Sentences 2 (LRS2) using only a fraction of the total labelled data.

In this work, we present the Densely Connected Temporal Convolutional Network (DC-TCN) for lip-reading of isolated words. Although Temporal Convolutional Networks (TCN) have recently demonstrated great potential in many vision tasks, its receptive fields are not dense enough to model the complex temporal dynamics in lip-reading scenarios. To address this problem, we introduce dense connections into the network to capture more robust temporal features. Moreover, our approach utilises the Squeeze-and-Excitation block, a light-weight attention mechanism, to further enhance the model's classification power. Without bells and whistles, our DC-TCN method has achieved 88.36% accuracy on the Lip Reading in the Wild (LRW) dataset and 43.65% on the LRW-1000 dataset, which has surpassed all the baseline methods and is the new state-of-the-art on both datasets.

Abstract Deep learning has catalysed progress in tasks such as face recognition and analysis, leading to a quick integration of technological solutions in multiple layers of our society. While such systems have proven to be accurate by standard evaluation metrics and benchmarks, a surge of work has recently exposed the demographic bias that such algorithms exhibit–highlighting that accuracy does not entail fairness. Clearly, deploying biased systems under real-world settings can have grave consequences for affected populations. Indeed, learning methods are prone to inheriting, or even amplifying the bias present in a training set, manifested by uneven representation across demographic groups. In facial datasets, this particularly relates to attributes such as skin tone, gender, and age. In this work, we address the problem of mitigating bias in facial datasets by data augmentation. We propose a multi-attribute framework that can successfully transfer complex, multi-scale facial patterns even if these belong to underrepresented groups in the training set. This is achieved by relaxing the rigid dependence on a single attribute label, and further introducing a tensor-based mixing structure that captures multiplicative interactions between attributes in a multilinear fashion. We evaluate our method with an extensive set of qualitative and quantitative experiments on several datasets, with rigorous comparisons to state-of-the-art methods. We find that the proposed framework can successfully mitigate dataset bias, as evinced by extensive evaluations on established diversity metrics, while significantly improving fairness metrics such as equality of opportunity.

Natural human-computer interaction and audio-visual human behaviour sensing systems, which would achieve robust performance in-the-wild are more needed than ever as digital devices are increasingly becoming an indispensable part of our life. Accurately annotated real-world data are the crux in devising such systems. However, existing databases usually consider controlled settings, low demographic variability, and a single task. In this paper, we introduce the SEWA database of more than 2,000 minutes of audio-visual data of 398 people coming from six cultures, 50 percent female, and uniformly spanning the age range of 18 to 65 years old. Subjects were recorded in two different contexts: while watching adverts and while discussing adverts in a video chat. The database includes rich annotations of the recordings in terms of facial landmarks, facial action units (FAU), various vocalisations, mirroring, and continuously valued valence, arousal, liking, agreement, and prototypic examples of (dis)liking. This database aims to be an extremely valuable resource for researchers in affective computing and automatic human sensing and is expected to push forward the research in human behaviour analysis, including cultural studies. Along with the database, we provide extensive baseline experiments for automatic FAU detection and automatic valence, arousal, and (dis)liking intensity estimation.

Predictability is important to autistic individuals, and robots have been suggested to meet this need as they can be programmed to be predictable, as well as elicit social interaction. The effectiveness of robot-assisted interventions designed for social skill learning presumably depends on the interplay between robot predictability, engagement in learning, and the individual differences between different autistic children. To better understand this interplay, we report on a study where 24 autistic children participated in a robot-assisted intervention. We manipulated the variance in the robot’s behaviour as a way to vary predictability, and measured the children’s behavioural engagement, visual attention, as well as their individual factors. We found that the children will continue engaging in the activity behaviourally, but may start to pay less visual attention over time to activity-relevant locations when the robot is less predictable. Instead, they increasingly start to look away from the activity. Ultimately, this could negatively influence learning, in particular for tasks with a visual component. Furthermore, severity of autistic features and expressive language ability had a significant impact on behavioural engagement. We consider our results as preliminary evidence that robot predictability is an important factor for keeping children in a state where learning can occur.

The prominence of deep learning, large amount of annotated data and increasingly powerful hardware made it possible to reach remarkable performance for supervised classification tasks, in many cases saturating the training sets. However the resulting models are specialized to a single very specific task and domain. Adapting the learned classification to new domains is a hard problem due to at least three reasons: (1) the new domains and the tasks might be drastically different; (2) there might be very limited amount of annotated data on the new domain and (3) full training of a new model for each new task is prohibitive in terms of computation and memory, due to the sheer number of parameters of deep CNNs. In this paper, we present a method to learn new-domains and tasks incrementally, building on prior knowledge from already learned tasks and without catastrophic forgetting. We do so by jointly parametrizing weights across layers using low-rank Tucker structure. The core is task agnostic while a set of task specific factors are learnt on each new domain. We show that leveraging tensor structure enables better performance than simply using matrix operations. Joint tensor modelling also naturally leverages correlations across different layers. Compared with previous methods which have focused on adapting each layer separately, our approach results in more compact representations for each new task/domain. We apply the proposed method to the 10 datasets of the Visual Decathlon Challenge and show that our method offers on average about 7.5× reduction in number of parameters and competitive performance in terms of both classification accuracy and Decathlon score.

For real-time semantic video segmentation, most recent works utilised a dynamic framework with a key scheduler to make online key/non-key decisions. Some works used a fixed key scheduling policy, while others proposed adaptive key scheduling methods based on heuristic strategies, both of which may lead to suboptimal global performance. To overcome this limitation, we model the online key decision process in dynamic video segmentation as a deep reinforcement learning problem and learn an efficient and effective scheduling policy from expert information about decision history and from the process of maximising global return. Moreover, we study the application of dynamic video segmentation on face videos, a field that has not been investigated before. By evaluating on the 300VW dataset, we show that the performance of our reinforcement key scheduler outperforms that of various baselines in terms of both effective key selections and running speed. Further results on the Cityscapes dataset demonstrate that our proposed method can also generalise to other scenarios. To the best of our knowledge, this is the first work to use reinforcement learning for online key-frame decision in dynamic video segmentation, and also the first work on its application on face videos.

Training deep neural networks with spatio-temporal (i.e., 3D) or multidimensional convolutions of higher-order is computationally challenging due to millions of unknown parameters across dozens of layers. To alleviate this, one approach is to apply low-rank tensor decompositions to convolution kernels in order to compress the network and reduce its number of parameters. Alternatively, new convolutional blocks, such as MobileNet, can be directly designed for efficiency. In this paper, we unify these two approaches by proposing a tensor factorization framework for efficient multidimensional (separable) convolutions of higher-order. Interestingly, the proposed framework enables a novel higher-order transduction, allowing to train a network on a given domain (e.g., 2D images or N-dimensional data in general) and using transduction to generalize to higher-order data such as videos (or (N+K)--dimensional data in general), capturing for instance temporal dynamics while preserving the learnt spatial information. We apply the proposed methodology, coined CP-Higher-Order Convolution (HO-CPConv), to spatio-temporal facial emotion analysis. Most existing facial affect models focus on static imagery and discard all temporal information. This is due to the above-mentioned burden of training 3D convolutional nets and the lack of large bodies of video data annotated by experts. We address both issues with our proposed framework. Initial training is first done on static imagery before using transduction to generalize to the temporal domain. We demonstrate superior performance on three challenging large scale affect estimation datasets, AffectNet, SEWA, and AFEW-VA.

This paper is on highly accurate and highly efficient human pose estimation. Recent works based on Fully Convolutional Networks (FCNs) have demonstrated excellent results for this difficult problem. While residual connections within FCNs have proved to be quintessential for achieving high accuracy, we re-analyze this design choice in the context of improving both the accuracy and the efficiency over the state-of- the-art. In particular, we make the following contributions: (a) We propose gated skip connections with per-channel learnable parameters to control the data flow for each channel within the module within the macro-module. (b) We introduce a hybrid network that combines the HourGlass and U-Net architectures which minimizes the number of identity connections within the network and increases the performance for the same parameter budget. Our model achieves state-of-the-art results on the MPII and LSP datasets. In addition, with a reduction of 3× in model size and complexity, we show no decrease in performance when compared to the original HourGlass network.

Visual face tracking is one of the most important components for face analysis in mobile applications and in video surveillance systems. In this paper, we propose an efficient face tracker called FT-RCNN, short for Face Tracking with Region-based CNN, that is based on the Faster-RCNN framework. A simple yet effective tracking branch is proposed to enable the framework to jointly perform face detection and tracking. To address the problem of insufficient training data for face tracking, we propose a novel pairwise training strategy that enables us to train face tracker by leveraging existing face detection datasets, thus eliminating the need to collect and annotate video data specifically for face tracking. Furthermore, we devise a novel loss function, termed Pair-hard Triplet Cosine Loss, that employs a pair-hard triplet mining strategy to increase the discriminative power of our face tracker. We evaluated the proposed tracker on popular video face datasets including MobiFace, ChokePoint and YouTube Face. The results have shown that FT-RCNN outperforms state-of-the-art face trackers and runs at real-time speed. We plan to release the source code for reproducible results in the future.

Lip-reading models have been significantly improved recently thanks to powerful deep learning architectures. However, most works focused on frontal or near frontal views of the mouth. As a consequence, lip-reading performance seriously deteriorates in non-frontal mouth views. In this work, we present a framework for training pose-invariant lip-reading models on synthetic data instead of collecting and annotating non-frontal data which is costly and tedious. The proposed model significantly outperforms previous approaches on non-frontal views while retaining the superior performance on frontal and near frontal mouth views. Specifically, we propose to use a 3D Morphable Model (3DMM) to augment LRW, an existing large-scale but mostly frontal dataset, by generating synthetic facial data in arbitrary poses. The newly derived dataset, is used to train a state-of-the-art neural network for lip-reading. We conducted a cross-database experiment for isolated word recognition on the LRS2 dataset, and reported an absolute improvement of 2.55%. The benefit of the proposed approach becomes clearer in extreme poses where an absolute improvement of up to 20.64% over the baseline is achieved.

Speech-driven facial animation involves using a speech signal to generate realistic videos of talking faces. Recent deep learning approaches to facial synthesis rely on extracting low-dimensional representations and concatenating them, followed by a decoding step of the concatenated vector. This accounts for only first-order interactions of the features and ignores higher-order interactions. In this paper we propose a polynomial fusion layer that models the joint representation of the encodings by a higher-order polynomial, with the parameters modelled by a tensor decomposition. We demonstrate the suitability of this approach through experiments on generated videos evaluated on a range of metrics on video quality, audiovisual synchronisation and generation of blinks.

Lip-reading has attracted a lot of research attention lately thanks to advances in deep learning. The current state-of-the-art model for recognition of isolated words in-the-wild consists of a residual network and Bidirectional Gated Recurrent Unit (BGRU) layers. In this work, we address the limitations of this model and we propose changes which further improve its performance. Firstly, the BGRU layers are replaced with Temporal Convolutional Networks (TCN). Secondly, we greatly simplify the training procedure, which allows us to train the model in one single stage. Thirdly, we show that the current state-of-the-art methodology produces models that do not generalize well to variations on the sequence length, and we address this issue by proposing a variable-length augmentation. We present results on the largest publicly-available datasets for isolated word recognition in English and Mandarin, LRW and LRW1000, respectively. Our proposed model1results in an absolute improvement of 1.2% and 3.2%, respectively, in these datasets which is the new state-of-the-art performance.

Parsimonious modelling, including sparsity and low rankness, has becomes a cornerstone in modern machine learning and signal processing. However, these modelling techniques have limited capabity to learn from large-scale data, and often require some pre-defined parameters to define their optimization procedure. In this paper, we propose a novel method to design specific deep neural networks for sparse and low-rank models, where the network can learn a data-adaptive model from training data. In particular, we design differentiable network units for sparse and low-rank matrices. Each layer of the network represents one iteration of the optimization process of sparse and low-rank models. The effectiveness of the proposed method is evaluated in the task of audio-visual object localization. Experimental results indicate the superior performance of the proposed method over traditional sparse and low-rank models.

Self supervised representation learning has recently attracted a lot of research interest for both the audio and visual modalities. However, most works typically focus on a particular modality or feature alone and there has been very limited work that studies the interaction between the two modalities for learning self supervised representations. We propose a framework for learning audio representations guided by the visual modality in the context of audiovisual speech. We employ a generative audio-to-video training scheme in which we animate a still image corresponding to a given audio clip and optimize the generated video to be as close as possible to the real video of the speech segment. Through this process, the audio encoder network learns useful speech representations that we evaluate on emotion recognition and speech recognition. We achieve state of the art results for emotion recognition and competitive results for speech recognition. This demonstrates the potential of visual supervision for learning audio representations as a novel way for self-supervised learning which has not been explored in the past. The proposed unsupervised audio features can leverage a virtually unlimited amount of training data of unlabelled audiovisual speech and have a large number of potentially promising applications.

Deep generative models rely on their inductive bias to facilitate generalization, especially for problems with high dimensional data, like images. However, empirical studies have shown that variational autoencoders (VAE) and generative adversarial networks (GAN) lack the generalization ability that occurs naturally in human perception. For example, humans can visualize a woman smiling after only seeing a smiling man. On the contrary, the standard conditional VAE (cVAE) is unable to generate unseen attribute combinations. To this end, we extend cVAE by introducing a multilinear latent conditioning framework that captures the multiplicative interactions between the attributes. We implement two variants of our model and demonstrate their efficacy on MNIST, Fashion-MNIST and CelebA. Altogether, we design a novel conditioning framework that can be used with any architecture to synthesize unseen attribute combinations.

Semantic segmentation of eyes has long been a vital pre-processing step in many biometric applications. Majority of the works focus only on high resolution eye images, while little has been done to segment the eyes from low quality images in the wild. However, this is a particularly interesting and meaningful topic, as eyes play a crucial role in conveying the emotional state and mental well-being of a person. In this work, we take two steps toward solving this problem: (1) We collect and annotate a challenging eye segmentation dataset containing 8882 eye patches from 4461 facial images of different resolutions, illumination conditions and head poses; (2) We develop a novel eye segmentation method, Shape Constrained Network (SCN), that incorporates shape prior into the segmentation network training procedure. Specifically, we learn the shape prior from our dataset using VAE-GAN, and leverage the pre-trained encoder and discriminator to regularise the training of SegNet. To improve the accuracy and quality of predicted masks, we replace the loss of SegNet with three new losses: Intersection-over-Union (IoU) loss, shape discriminator loss and shape embedding loss. Extensive experiments shows that our method outperforms state-of-the-art segmentation and landmark detection methods in terms of mean IoU (mIoU) accuracy and the quality of segmentation masks. The dataset is available at https://ibug.doc.ic.ac.uk/resources/ibug-eye-segmentation-dataset/.

Abstract Speech-driven facial animation is the process that automatically synthesizes talking characters based on speech signals. The majority of work in this domain creates a mapping from audio features to visual features. This approach often requires post-processing using computer graphics techniques to produce realistic albeit subject dependent results. We present an end-to-end system that generates videos of a talking head, using only a still image of a person and an audio clip containing speech, without relying on handcrafted intermediate features. Our method generates videos which have (a) lip movements that are in sync with the audio and (b) natural facial expressions such as blinks and eyebrow movements. Our temporal GAN uses 3 discriminators focused on achieving detailed frames, audio-visual synchronization, and realistic expressions. We quantify the contribution of each component in our model using an ablation study and we provide insights into the latent representation of the model. The generated videos are evaluated based on sharpness, reconstruction quality, lip-reading accuracy, synchronization as well as their ability to generate natural blinks.

A novel method for active speaker detection and localization in audio-visual recordings is proposed. The method relies on a specifically tailored matrix decomposition that exploits the intrinsic low-dimensional structure of audio-visual data, namely, the low-rank of the background visual/audio information and the sparsity of the correlated foreground components. Concretely, the data matrix of each modality is modeled as a superposition of two terms: 1) a low-rank matrix capturing the background information and 2) a kernelized sparse matrix capturing the non-linear correlated components among the audio and visual modalities and, hence, revealing the active speaker. To this end, we formulate an appropriate optimization problem that involves the minimization of nuclear- and matrix $\ell _1$-norms, and develop an efficient solver. Experimental results on active speaker detection and localization demonstrate the superior performance of the proposed method over other state-of-the-art approaches.

The ability to localize visual objects that are associated with an audio source and at the same time to separate the audio signal is a cornerstone in audio-visual signal-processing applications. However, available methods mainly focus on localizing only the visual objects, without audio separation abilities. Besides that, these methods often rely on either laborious preprocessing steps to segment video frames into semantic regions, or additional supervisions to guide their localization. In this paper, we aim to address the problem of visual source localization and audio separation in an unsupervised manner and avoid all preprocessing or post-processing steps. To this end, we devise a novel structured matrix decomposition method that decomposes the data matrix of each modality as a superposition of three terms: 1) a low-rank matrix capturing the background information; 2) a sparse matrix capturing the correlated components among the two modalities and, hence, uncovering the sound source in visual modality and the associated sound in audio modality; and 3) a third sparse matrix accounting for uncorrelated components, such as distracting objects in visual modality and irrelevant sound in audio modality. The generality of the proposed method is demonstrated by applying it onto three applications, namely: 1) visual localization of a sound source; 2) visually assisted audio separation; and 3) active speaker detection. Experimental results indicate the effectiveness of the proposed method on these application domains.

Recent findings indicate that over-parametrization, while crucial for successfully training deep neural networks, also introduces large amounts of redundancy. Tensor methods have the potential to efficiently parametrize over-complete representations by leveraging this redundancy. In this paper, we propose to fully parametrize Convolutional Neural Networks (CNNs) with a single high-order, low-rank tensor. Previous works on network tensorization have focused on parametrizing individual layers (convolutional or fully connected) only, and perform the tensorization layer-by-layer separately. In contrast, we propose to jointly capture the full structure of a neural network by parametrizing it with a single high-order tensor, the modes of which represent each of the architectural design parameters of the network (e.g. number of convolutional blocks, depth, number of stacks, input features, etc). This parametrization allows to regularize the whole network and drastically reduce the number of parameters. Our model is end-to-end trainable and the low-rank structure imposed on the weight tensor acts as an implicit regularization. We study the case of networks with rich structure, namely Fully Convolutional Networks (FCNs), which we propose to parametrize with a single 8th-order tensor. We show that our approach can achieve superior performance with small compression rates, and attain high compression rates with negligible drop in accuracy for the challenging task of human pose estimation.

Face tracking serves as the crucial initial step in mobile applications trying to analyse target faces over time in mobile settings. However, this problem has received little attention, mainly due to the scarcity of dedicated face tracking benchmarks. In this work, we introduce MobiFace, the first dataset for single face tracking in mobile situations. It consists of 80 unedited live-streaming mobile videos captured by 70 different smartphone users in fully unconstrained environments. Over 95K bounding boxes are manually labelled. The videos are carefully selected to cover typical smartphone usage. The videos are also annotated with 14 attributes, including 6 newly proposed attributes and 8 commonly seen in object tracking. 36 state-of-the-art trackers, including facial landmark trackers, generic object trackers and trackers that we have fine-tuned or improved, are evaluated. The results suggest that mobile face tracking cannot be solved through existing approaches. In addition, we show that fine-tuning on the MobiFace training data significantly boosts the performance of deep learning-based trackers, suggesting that MobiFace captures the unique characteristics of mobile face tracking. Our goal is to offer the community a diverse dataset to enable the design and evaluation of mobile face trackers. The dataset, annotations and the evaluation server will be on https://mobiface.github.io/.

While it is relatively easy and natural for humans to detect and interpret non-verbal cues, it is a hard task for computer systems. Automatic recognition of emotions has been the subject of extensive studies in the past decade, but despite the various methods that have been implemented, the problem remains challenging. In particular, most existing works focus on predicting a set discrete stereotypical emotion categories. We are instead interested in predicting continuous values of valence and arousal, which are able to model accurately a broader range of spontaneous emotions. Moreover, as opposed to much of the prior work that focused on controlled (laboratory) conditions, we are interested in analysis in naturalist (in-the-wild) conditions. To do so, we propose to leverage the structure in the data using tensor methods. In addition to preserving the structure, these have the potential to also reduce the total number of parameters in the models, thus improving the computational performance. We first consider a model with analytic solution in the form of a Tucker Tensor Regression. We then investigate a deep, gradient based method, namely Tensor Regression Networks. We perform thorough experiments on two publicly available databases, AFEW-VA and SEWA, for facial affect estimation in-the-wild, in terms of valence and arousal levels. Experimental results demonstrate that tensor-based methods successfully leverage the structure in the data and on average outperform baseline methods.

Robust principal component analysis (RPCA), decomposes a data matrix into a superposition of a low-rank matrix and a sparse matrix under certain incoherent conditions. In this paper, we propose a nonlinear generalization of RPCA that uses two autoencoder networks to achieve such a decomposition, in which one autoencoder accounts for the low-rank component and the other for the sparse component. To this end, we provide a principled way of constructing these autoencoders for low-rank and sparse components. The generality of the proposed model is demonstrated by applying it onto three applications, namely 1) music/voice separation 2) image denoising and 3) video foreground separation. Experimental results indicate the effectiveness of the proposed model on these application domains.

Several audio-visual speech recognition models have been recently proposed which aim to improve the robustness over audio-only models in the presence of noise. However, almost all of them ignore the impact of the Lombard effect, i.e., the change in speaking style in noisy environments which aims to make speech more intelligible and affects both the acoustic characteristics of speech and the lip movements. In this paper, we investigate the impact of the Lombard effect in audio-visual speech recognition. To the best of our knowledge, this is the first work which does so using end-to-end deep architectures and presents results on unseen speakers. Our results show that properly modelling Lombard speech is always beneficial. Even if a relatively small amount of Lombard speech is added to the training set then the performance in a real scenario, where noisy Lombard speech is present, can be significantly improved. We also show that the standard approach followed in the literature, where a model is trained and tested on noisy plain speech, provides a correct estimate of the video-only performance and slightly underestimates the audio-visual performance. In case of audio-only approaches, performance is overestimated for SNRs higher than -3dB and underestimated for lower SNRs.

Speech is a means of communication which relies on both audio and visual information. The absence of one modality can often lead to confusion or misinterpretation of information. In this paper we present an end-to-end temporal model capable of directly synthesising audio from silent video, without needing to transform to-and-from intermediate features. Our proposed approach, based on GANs is capable of producing natural sounding, intelligible speech which is synchronised with the video. The performance of our model is evaluated on the GRID dataset for both speaker dependent and speaker independent scenarios. To the best of our knowledge this is the first method that maps video directly to raw audio and the first to produce intelligible speech when tested on previously unseen speakers. We evaluate the synthesised audio not only based on the sound quality but also on the accuracy of the spoken words.

The ninth Audio-Visual Emotion Challenge and workshop AVEC 2019 was held in conjunction with ACM Multimedia'19. This year, the AVEC series addressed major novelties with three distinct tasks: State-of-Mind Sub-challenge (SoMS), Detecting Depression with Artificial Intelligence Sub-challenge (DDS), and Cross-cultural Emotion Sub-challenge (CES). The SoMS was based on a novel dataset (USoM corpus) that includes self-reported mood (10-point Likert scale) after the narrative of personal stories (two positive and two negative). The DDS was based on a large extension of the DAIC-WOZ corpus (c.f. AVEC 2016) that includes new recordings of patients suffering from depression with the virtual agent conducting the interview being, this time, wholly driven by AI, i.e., without any human intervention. The CES was based on the SEWA dataset (c.f. AVEC 2018) that has been extended with the inclusion of new participants in order to investigate how emotion knowledge of Western European cultures (German, Hungarian) can be transferred to the Chinese culture. In this summary, we mainly describe participation and conditions of the AVEC Challenge.

Deep generative models have significantly advanced image generation, enabling generation of visually pleasing images with realistic texture. Apart from the texture, it is the shape geometry of objects that strongly dictates their appearance. However, currently available generative models do not incorporate geometric information into the image generation process. This often yields visual objects of degenerated quality. In this work, we propose a regularized Geometry-Aware Generative Adversarial Network (GAGAN) which disentangles appearance and shape in the latent space. This regularized GAGAN enables the generation of images with both realistic texture and shape. Specifically, we condition the generator on a statistical shape prior. The prior is enforced through mapping the generated images onto a canonical coordinate frame using a differentiable geometric transformation. In addition to incorporating geometric information, this constrains the search space and increases the model’s robustness. We show that our approach is versatile, able to generalise across domains (faces, sketches, hands and cats) and sample sizes (from as little as $$\sim \, 200{-}30{,}000$$ to more than 200, 000). We demonstrate superior performance through extensive quantitative and qualitative experiments in a variety of tasks and settings. Finally, we leverage our model to automatically and accurately detect errors or drifting in facial landmarks detection and tracking in-the-wild.

Inspired by the recent development of deep network-based methods in semantic image segmentation, we introduce an end-to-end trainable model for face mask extraction in video sequence. Comparing to landmark-based sparse face shape representation, our method can produce the segmentation masks of individual facial components, which can better reflect their detailed shape variations. By integrating convolutional LSTM (ConvLSTM) algorithm with fully convolutional networks (FCN), our new ConvLSTM-FCN model works on a per-sequence basis and takes advantage of the temporal correlation in video clips. In addition, we also propose a novel loss function, called segmentation loss, to directly optimise the intersection over union (IoU) performances. In practice, to further increase segmentation accuracy, one primary model and two additional models were trained to focus on the face, eyes, and mouth regions, respectively. Our experiment shows the proposed method has achieved a 16.99% relative improvement (from 54.50 to 63.76% mean IoU) over the baseline FCN model on the 300 Videos in the Wild (300VW) dataset.

Tensors are higher-order extensions of matrices. While matrix methods form the cornerstone of traditional machine learning and data analysis, tensor methods have been gaining increasing traction. However, software support for tensor operations is not on the same footing. In order to bridge this gap, we have developed TensorLy, a Python library that provides a high-level API for tensor methods and deep tensorized neural networks. TensorLy aims to follow the same standards adopted by the main projects of the Python scientific community, and to seamlessly integrate with them. Its BSD license makes it suitable for both academic and commercial applications. TensorLy's backend system allows users to perform computations with several libraries such as NumPy or PyTorch to name but a few. They can be scaled on multiple CPU or GPU machines. In addition, using the deep-learning frameworks as backend allows to easily design and train deep tensorized neural networks. TensorLy is available at https://github.com/tensorly/tensorly

As one of the most comprehensive and objective ways to describe facial expressions, the Facial Action Coding System (FACS) has recently received significant attention. Over the past 30 years, extensive research has been conducted by psychologists and neuroscientists on various aspects of facial expression analysis using FACS. Automating FACS coding would make this research faster and more widely applicable, opening up new avenues to understanding how we communicate through facial expressions. Such an automated process can also potentially increase the reliability, precision and temporal resolution of coding. This paper provides a comprehensive survey of research into machine analysis of facial actions. We systematically review all components of such systems: pre-processing, feature extraction and machine coding of facial actions. In addition, the existing FACS-coded facial expression databases are summarised. Finally, challenges that have to be addressed to make automatic facial action analysis applicable in real-life situations are extensively discussed. There are two underlying motivations for us to write this survey paper: the first is to provide an up-to-date review of the existing literature, and the second is to offer some insights into the future of machine recognition of facial actions: what are the challenges and opportunities that researchers in the field face.

Joint modeling of the intensity of multiple facial action units (AUs) from face images is challenging due to the large number of AUs (30+) and their intensity levels (6). This is in part due to the lack of suitable models that can efficiently handle such a large number of outputs/classes simultaneously, but also due to the lack of suitable data the models on. For this reason, majority of the methods resort to independent classifiers for the AU intensity. This is suboptimal for at least two reasons: the facial appearance of some AUs changes depending on the intensity of other AUs, and some AUs co-occur more often than others. To this end, we propose the Copula regression approach for modeling multivariate ordinal variables. Our model accounts for ordinal structure in output variables and their non-linear dependencies via copula functions modeled as cliques of a conditional random fields. The copula ordinal regression model achieves the joint learning and inference of intensities of multiple AUs, while being computationally tractable. We demonstrate the effectiveness of our approach on three challenging datasets of naturalistic facial expressions and we show that the estimation of target AU intensities improves especially in the case of (a) noisy image features, (b) head-pose variations and (c) imbalanced training data. Lastly, we show that the proposed approach consistently outperforms (i) independent modeling of AU intensities and (ii) the state-of-the-art approach for the target task and (iii) deep convolutional neural networks.

The progress we are currently witnessing in many computer vision applications, including automatic face analysis, would not be made possible without tremendous efforts in collecting and annotating large scale visual databases. To this end, we propose 4DFAB, a new large scale database of dynamic high-resolution 3D faces (over 1,800,000 3D meshes). 4DFAB contains recordings of 180 subjects captured in four different sessions spanning over a five-year period. It contains 4D videos of subjects displaying both spontaneous and posed facial behaviours. The database can be used for both face and facial expression recognition, as well as behavioural biometrics. It can also be used to learn very powerful blendshapes for parametrising facial behaviour. In this paper, we conduct several experiments and demonstrate the usefulness of the database for various applications. The database will be made publicly available for research purposes.

Deep generative models learned through adversarial training have become increasingly popular for their ability to generate naturalistic image textures. However, aside from their texture, the visual appearance of objects is significantly influenced by their shape geometry; information which is not taken into account by existing generative models. This paper introduces the Geometry-Aware Generative Adversarial Networks (GAGAN) for incorporating geometric information into the image generation process. Specifically, in GAGAN the generator samples latent variables from the probability space of a statistical shape model. By mapping the output of the generator to a canonical coordinate frame through a differentiable geometric transformation, we enforce the geometry of the objects and add an implicit connection from the prior to the generated object. Experimental results on face generation indicate that the GAGAN can generate realistic images of faces with arbitrary facial attributes such as facial expression, pose, and morphology, that are of better quality than current GAN-based methods. Our method can be used to augment any existing GAN architecture and improve the quality of the images generated.

Conflict arises naturally in dyadic interactions when involved individuals act on incompatible goals, interests, or actions. In this paper, the problem of conflict intensity estimation from audiovisual recordings is addressed. To this end, we propose an online attention-based neural network in order to learn a mapping from a sequence of audiovisual features to time-series describing conflict intensity. The proposed method is evaluated by conducting experiments in conflict intensity estimation by employing the CONFER dataset. Experimental results indicate the superiority of the proposed model compared to the state of the art. Furthermore, we demonstrate that by incorporating sparsity in the model, the origin of conflict can be traced back to specific key frames facilitating the interpretation of conflict escalation.

Silent speech interfaces have been recently proposed as a way to enable communication when the acoustic signal is not available. This introduces the need to build visual speech recognition systems for silent and whispered speech. However, almost all the recently proposed systems have been trained on vocalised data only. This is in contrast with evidence in the literature which suggests that lip movements change depending on the speech mode. In this work, we introduce a new audiovisual database which is publicly available and contains normal, whispered and silent speech. To the best of our knowledge, this is the first study which investigates the differences between the three speech modes using the visual modality only. We show that an absolute decrease in classification rate of up to 3.7% is observed when training and testing on normal and whispered, respectively, and vice versa. An even higher decrease of up to 8.5% is reported when the models are tested on silent speech. This reveals that there are indeed visual differences between the 3 speech modes and the common assumption that vocalized training data can be used directly to train a silent speech recognition system may not be true.

Several end-to-end deep learning approaches have been recently presented which extract either audio or visual features from the input images or audio signals and perform speech recognition. However, research on end-to-end audiovisual models is very limited. In this work, we present an end-to-end audiovisual model based on residual networks and Bidirectional Gated Recurrent Units (BGRUs). To the best of our knowledge, this is the first audiovisual fusion model which simultaneously learns to extract features directly from the image pixels and audio waveforms and performs within-context word recognition on a large publicly available dataset (LRW). The model consists of two streams, one for each modality, which extract features directly from mouth regions and raw waveforms. The temporal dynamics in each stream/modality are modeled by a 2-layer BGRU and the fusion of multiple streams/modalities takes place via another 2-layer BGRU. A slight improvement in the classification rate over an end-to-end audio-only and MFCC-based model is reported in clean audio conditions and low levels of noise. In presence of high levels of noise, the end-to-end audiovisual model significantly outperforms both audio-only models.

The eighth Audio-Visual Emotion Challenge and workshop AVEC 2018 was held in conjunction with ACM Multimedia'18. This year, the AVEC series addressed major novelties with three distinct sub-challenges: bipolar disorder classification, cross-cultural dimensional emotion recognition, and emotional label generation from individual ratings. The Bipolar Disorder Sub-challenge was based on a novel dataset of structured interviews of patients suffering from bipolar disorder (BD corpus), the Cross-cultural Emotion Sub-challenge relied on an extension of the SEWA dataset, which includes human-human interactions recorded 'in-the-wild' for the German and the Hungarian cultures, and the Gold-standard Emotion Sub-challenge was based on the RECOLA dataset, which was previously used in the AVEC series for emotion recognition. In this summary, we mainly describe participation and conditions of the AVEC Challenge.

Recent works in speech recognition rely either on connectionist temporal classification (CTC) or sequence-to-sequence models for character-level recognition. CTC assumes conditional independence of individual characters, whereas attention-based models can provide nonsequential alignments. Therefore, we could use a CTC loss in combination with an attention-based model in order to force monotonic alignments and at the same time get rid of the conditional independence assumption. In this paper, we use the recently proposed hybrid CTC/attention architecture for audio-visual recognition of speech in-the-wild. To the best of our knowledge, this is the first time that such a hybrid architecture architecture is used for audio-visual recognition of speech. We use the LRS2 database and show that the proposed audio-visual model leads to an 1.3% absolute decrease in word error rate over the audio-only model and achieves the new state-of-the-art performance on LRS2 database (7% word error rate). We also observe that the audio-visual model significantly outperforms the audio-based model (up to 32.9% absolute improvement in word error rate) for several different types of noise as the signal-to-noise ratio decreases.

Human behavior and affect is inherently a dynamic phenomenon involving temporal evolution of patterns manifested through a multiplicity of non-verbal behavioral cues including facial expressions, body postures and gestures, and vocal outbursts. A natural assumption for human behavior modeling is that a continuous-time characterization of behavior is the output of a linear time-invariant system when behavioral cues act as the input (e.g., continuous rather than discrete annotations of dimensional affect). Here we study the learning of such dynamical system under real-world conditions, namely in the presence of noisy behavioral cues descriptors and possibly unreliable annotations by employing structured rank minimization. To this end, a novel structured rank minimization method and its scalable variant are proposed. The generalizability of the proposed framework is demonstrated by conducting experiments on 3 distinct dynamic behavior analysis tasks, namely (i) conflict intensity prediction, (ii) prediction of valence and arousal, and (iii) tracklet matching. The attained results outperform those achieved by other state-of-the-art methods for these tasks and, hence, evidence the robustness and effectiveness of the proposed approach.

The papers in this special section examine the concept of automated face analysis (AFA). AFA has received special attention from the computer vision and pattern recognition communities. Research progress often gives the impression that problems such as face recognition and face detection are solved, at least for some scenarios. Several aspects of face analysis remain open problems, including the implementation of large scale face recognition/detection methods for in the wild images, emotion recognition, micro-expression analysis, and others. The community keeps making rapid progress on these topics, with continual improvement of current methods and creation of new ones that push the state-of-the-art. Applications are countless, including security and video surveillance, human computer/robot interaction, communication, entertainment, and commerce, while having an important social impact in assistive technologies for education and health. The importance of face analysis, together with the vast amount of work on the subject and the latest developments in the field, motivated us to organize a special section on this theme. The scope of the compilation comprises all aspects of face analysis from a computer vision perspective. Including, but not limited to: recognition, detection, alignment, reconstruction of faces, pose estimation of faces, gaze analysis, age, emotion, gender, and facial attributes estimation, and applications among others.

We propose a Multi-Instance-Learning (MIL) approach for weakly-supervised learning problems, where a training set is formed by bags (sets of feature vectors or instances) and only labels at bag-level are provided. Specifically, we consider the Multi-Instance Dynamic-Ordinal-Regression (MI-DOR) setting, where the instance labels are naturally represented as ordinal variables and bags are structured as temporal sequences. To this end, we propose Multi-Instance Dynamic Ordinal Random Fields (MI-DORF). In this framework, we treat instance-labels as temporally-dependent latent variables in an Undirected Graphical Model. Different MIL assumptions are modelled via newly introduced high-order potentials relating bag and instance-labels within the energy function of the model. We also extend our framework to address the Partially-Observed MI-DOR problems, where a subset of instance labels are available during training.We show on the tasks of weakly-supervised facial behavior analysis, Facial Action Unit (DISFA dataset) and Pain (UNBC dataset) Intensity estimation, that the proposed framework outperforms alternative learning approaches. Furthermore, we show that MIDORF can be employed to reduce the data annotation efforts in this context by large-scale.

No abstract available.

We consider the task of automated estimation of facial expression intensity. This involves estimation of multiple output variables (facial action units - AUs) that are structurally dependent. Their structure arises from statistically induced co-occurrence patterns of AU intensity levels. Modeling this structure is critical for improving the estimation performance, however, this performance is bounded by the quality of the input features extracted from face images. The goal of this paper is to model these structures and estimate complex feature representations simultaneously by combining conditional random field (CRF) encoded AU dependencies with deep learning. To this end, we propose a novel Copula CNN deep learning approach for modeling multivariate ordinal variables. Our model accounts for ordinal structure in output variables and their non-linear dependencies via copula functions modeled as cliques of a CRF. These are jointly optimized with deep CNN feature encoding layers using a newly introduced balanced batch iterative training algorithm. We demonstrate the effectiveness of our approach on the task of AU intensity estimation on two benchmark datasets. We show that joint learning of the deep features and the target output structure results in significant performance gains compared to existing structured deep models and deep models for analysis of facial expressions.

Unsupervised learning of invariant representations that efficiently describe high-dimensional time series has several applications in dynamic visual data analysis. Clearly, the problem becomes more challenging when dealing with multiple time series arising from different modalities. A prominent example of this multimodal setting is the human motion which can be represented by multimodal time series of pixel intensities, depth maps, and motion capture data. Here, we study, for the first time, the problem of unsupervised learning of temporally and modality invariant informative representations, referred to as archetypes, from multiple time series originating from different modalities. To this end a novel method, coined as temporal archetypal analysis, is proposed. The performance of the proposed method is assessed by conducting experiments in unsupervised action segmentation. Experimental results on three different real world datasets using single modal and multimodal visual representations indicate the robustness and effectiveness of the proposed methods, outperforming compared state-of-the-art methods by a large, in most of the cases, margin.

Discriminative classification models have been successfully applied for various computer vision tasks such as object and face detection and recognition. However, deformations can change objects coordinate space and perturb robust similarity measurement, which is the essence of all classification algorithms. The common approach to deal with deformations is either to seek for deformation invariant features or to develop models that describe objects deformations. However, the former approach requires a huge amount of data and a good amount of engineering to be properly trained, while the latter require considerable human effort in the form of carefully annotated data. In this paper, we propose a method that jointly learns with minimal human intervention a generative deformable model using only a simple shape model of the object and images automatically downloaded from the Internet, and also extracts features appropriate for classification. The proposed algorithm is applied on various classification problems such as “in-thewild” face recognition, gender classification and eye glasses detection on data retrieved by querying into a web image search engine. We demonstrate that not only it outperforms other automatic methods by large margins, but also performs comparably with supervised methods trained on thousands of manually annotated data.

The field of Automatic Facial Expression Analysis has grown rapidly in recent years. However, despite progress in new approaches as well as benchmarking efforts, most evaluations still focus on either posed expressions, near-frontal recordings, or both. This makes it hard to tell how existing expression recognition approaches perform under conditions where faces appear in a wide range of poses (or camera views), displaying ecologically valid expressions. The main obstacle for assessing this is the availability of suitable data, and the challenge proposed here addresses this limitation. The FG 2017 Facial Expression Recognition and Analysis challenge (FERA 2017) extends FERA 2015 to the estimation of Action Units occurrence and intensity under different camera views. In this paper we present the third challenge in automatic recognition of facial expressions, to be held in conjunction with the 12th IEEE conference on Face and Gesture Recognition, May 2017, in Washington, United States. Two sub-challenges are defined: the detection of AU occurrence, and the estimation of AU intensity. In this work we outline the evaluation protocol, the data used, and the results of a baseline method for both sub-challenges.

Traditional visual speech recognition systems consist of two stages, feature extraction and classification. Recently, several deep learning approaches have been presented which automatically extract features from the mouth images and aim to replace the feature extraction stage. However, research on joint learning of features and classification is very limited. In this work, we present an end-to-end visual speech recognition system based on Long-Short Memory (LSTM) networks. To the best of our knowledge, this is the first model which simultaneously learns to extract features directly from the pixels and perform classification and also achieves state-of-the-art performance in visual speech classification. The model consists of two streams which extract features directly from the mouth and difference images, respectively. The temporal dynamics in each stream are modelled by an LSTM and the fusion of the two streams takes place via a Bidirectional LSTM (BLSTM). An absolute improvement of 9.7% over the base line is reported on the OuluVS2 database, and 1.5% on the CUAVE database when compared with other methods which use a similar visual front-end.

The ability to localize visual objects that are associated with an audio source and at the same time seperate the audio signal is a corner stone in several audio-visual signal processing applications. Past efforts usually focused on localizing only the visual objects, without audio separation abilities. Besides, they often rely computational expensive pre-processing steps to segment images pixels into object regions before applying localization approaches. We aim to address the problem of audio-visual source localization and separation in an unsupervised manner. The proposed approach employs low-rank in order to model the background visual and audio information and sparsity in order to extract the sparsely correlated components between the audio and visual modalities. In particular, this model decomposes each dataset into a sum of two terms: the low-rank matrices capturing the background uncorrelated information, while the sparse correlated components modelling the sound source in visual modality and the associated sound in audio modality. To this end a novel optimization problem, involving the minimization of nuclear norms and matrix ℓ1-norms is solved. We evaluated the proposed method in 1) visual localization and audio separation and 2) visual-assisted audio denoising. The experimental results demonstrate the effectiveness of the proposed method.

Human face exhibits an inherent hierarchy in its representations (i.e., holistic facial expressions can be encoded via a set of facial action units (AUs) and their intensity). Variational (deep) auto-encoders (VAE) have shown great results in unsupervised extraction of hierarchical latent representations from large amounts of image data, while being robust to noise and other undesired artifacts. Potentially, this makes VAEs a suitable approach for learning facial features for AU intensity estimation. Yet, most existing VAE-based methods apply classifiers learned separately from the encoded features. By contrast, the non-parametric. (probabilistic) approaches, such as Gaussian Processes (GPs), typically outperform their parametric counterparts, but cannot deal easily with large amounts of data. To this end, we propose a novel VAE semi-parametric modeling framework, named DeepCoder, which combines the modeling power of parametric (convolutional) and non-parametric. (ordinal GPs) VAEs, for joint learning of(l) latent representations at multiple levels in a task hierarchy1, and (2) classification of multiple ordinal outputs. We show on benchmark datasets for AU intensity estimation that the proposed DeepCoder outperforms the state-of-the-art approaches, and related VAEs and deep learning models.

The automatic detection and classification of social signals is an important task, given the fundamental role nonverbal behavioral cues play in human communication. We present the first cross-lingual study on the detection of laughter and fillers in conversational and spontaneous speech collected 'in the wild' over IP (internet protocol). Further, this is the first comparison of LSTM and GRU networks to shed light on their performance differences. We report frame-based results in terms of the unweighted-average area-under-the-curve (UAAUC) measure and will shortly discuss its suitability for this task. In the mono-lingual setup our best deep BLSTM system achieves 87.0% and 86.3% UAAUC for English and German, respectively. Interestingly, the cross-lingual results are only slightly lower, yielding 83.7% for a system trained on English, but tested on German, and 85.0% in the opposite case. We show that LSTM and GRU architectures are valid alternatives for e. g., on-line and compute-sensitive applications, since their application incurs a relative UAAUC decrease of only approximately 5% with respect to our best systems. Finally, we apply additional smoothing to correct for erroneous spikes and drops in the posterior trajectories to obtain an additional gain in all setups.

Over the last decade, automatic emotion recognition has become well established. The gold standard target is thereby usually calculated based on multiple annotations from different raters. All related efforts assume that the emotional state of a human subject can be identified by a 'hard' category or a unique value. This assumption tries to ease the human observer's subjectivity when observing patterns such as the emotional state of others. However, as the number of annotators cannot be infinite, uncertainty remains in the emotion target even if calculated from several, yet few human annotators. The common procedure to use this same emotion target in the learning process thus inevitably introduces noise in terms of an uncertain learning target. In this light, we propose a 'soft' prediction framework to provide a more human-like and comprehensive prediction of emotion. In our novel framework, we provide an additional target to indicate the uncertainty of human perception based on the inter-rater disagreement level, in contrast to the traditional framework which is merely producing one single prediction (category or value). To exploit the dependency between the emotional state and the newly introduced perception uncertainty, we implement a multi-task learning strategy. To evaluate the feasibility and effectiveness of the proposed soft prediction framework, we perform extensive experiments on a time- and value-continuous spontaneous audiovisual emotion database including late fusion results. We show that the soft prediction framework with multi-task learning of the emotional state and its perception uncertainty significantly outperforms the individual tasks in both the arousal and valence dimensions.

The seventh Audio-Visual Emotion Challenge and workshop AVEC 2017 was held in conjunction with ACM Multimedia'17. This year, the AVEC series addresses two distinct sub-challenges: emotion recognition and depression detection. The Affect Sub-Challenge is based on a novel dataset of human-human interactions recorded 'in-the-wild', whereas the Depression Sub-Challenge is based on the same dataset as the one used in AVEC 2016, with human-agent interactions. In this summary, we mainly describe participation and its conditions.

The unconstrained acquisition of facial data in real-world conditions may result in face images with significant pose variations, illumination changes, and occlusions, affecting the performance of facial landmark localization and recognition methods. In this paper, a novel method, robust to pose, illumination variations, and occlusions is proposed for joint face frontalization and landmark localization. Unlike the state-of-the-art methods for landmark localization and pose correction, where large amount of manually annotated images or 3D facial models are required, the proposed method relies on a small set of frontal images only. By observing that the frontal facial image of both humans and animals, is the one having the minimum rank of all different poses, a model which is able to jointly recover the frontalized version of the face as well as the facial landmarks is devised. To this end, a suitable optimization problem is solved, concerning minimization of the nuclear norm (convex surrogate of the rank function) and the matrix $$\ell _1$$ norm accounting for occlusions. The proposed method is assessed in frontal view reconstruction of human and animal faces, landmark localization, pose-invariant face recognition, face verification in unconstrained conditions, and video inpainting by conducting experiment on 9 databases. The experimental results demonstrate the effectiveness of the proposed method in comparison to the state-of-the-art methods for the target problems.

Fitting algorithms for Active Appearance Models (AAMs) are usually considered to be robust but slow or fast but less able to generalize well to unseen variations. In this paper, we look into AAM fitting algorithms and make the following orthogonal contributions: We present a simple "project-out" optimization framework that unifies and revises the most well-known optimization problems and solutions in AAMs. Based on this framework, we describe robust simultaneous AAM fitting algorithms the complexity of which is not prohibitive for current systems. We then go on one step further and propose a new approximate project-out AAM fitting algorithm which we coin Extended Project-Out Inverse Compositional (E-POIC). In contrast to current algorithms, E-POIC is both efficient and robust. Next, we describe a part-based AAM employing a translational motion model, which results in superior fitting and convergence properties. We also show that the proposed AAMs, when trained "in-the-wild" using SIFT descriptors, perform surprisingly well even for the case of unseen unconstrained images. Via a number of experiments on unconstrained human and animal face databases, we show that our combined contributions largely bridge the gap between exact and current approximate methods for AAM fitting and perform comparably with state-of-the-art face alignment systems.

Most of existing models for facial behavior analysis rely on generic classifiers, which fail to generalize well to previously unseen data. This is because of inherent differences in source (training) and target (test) data, mainly caused by variation in subjects' facial morphology, camera views, and so on. All of these account for different contexts in which target and source data are recorded, and thus, may adversely affect the performance of the models learned solely from source data. In this paper, we exploit the notion of domain adaptation and propose a data efficient approach to adapt already learned classifiers to new unseen contexts. Specifically, we build upon the probabilistic framework of Gaussian processes (GPs), and introduce domain-specific GP experts (e.g., for each subject). The model adaptation is facilitated in a probabilistic fashion, by conditioning the target expert on the predictions from multiple source experts. We further exploit the predictive variance of each expert to define an optimal weighting during inference. We evaluate the proposed model on three publicly available data sets for multi-class (MultiPIE) and multi-label (DISFA, FERA2015) facial expression analysis by performing adaptation of two contextual factors: "where" (view) and "who" (subject). In our experiments, the proposed approach consistently outperforms: 1) both source and target classifiers, while using a small number of target examples during the adaptation and 2) related state-of-the-art approaches for supervised domain adaptation.

Active appearance models (AAMs) are generative models of shape and appearance that have proven very attractive for their ability to handle wide changes in illumination, pose, and occlusion when trained in the wild, while not requiring large training data set like regression-based or deep learning methods. The problem of fitting an AAM is usually formulated as a non-linear least squares one and the main way of solving it is a standard Gauss-Newton algorithm. In this paper, we extend AAMs in two ways: we first extend the Gauss-Newton framework by formulating a bidirectional fitting method that deforms both the image and the template to fit a new instance. We then formulate a second order method by deriving an efficient Newton method for AAMs fitting. We derive both methods in a unified framework for two types of AAMs, holistic and part-based, and additionally show how to exploit the structure in the problem to derive fast yet exact solutions. We perform a thorough evaluation of all algorithms on three challenging and recently annotated in-the-wild data sets, and investigate fitting accuracy, convergence properties, and the influence of noise in the initialization. We compare our proposed methods to other algorithms and show that they yield state-of-the-art results, out-performing other methods while having superior convergence properties.

The analysis of high-dimensional, possibly temporally misaligned, and time-varying visual data is a fundamental task in disciplines, such as image, vision, and behavior computing. In this paper, we focus on dynamic facial behavior analysis and in particular on the analysis of facial expressions. Distinct from the previous approaches, where sets of facial landmarks are used for face representation, raw pixel intensities are exploited for: 1) unsupervised analysis of the temporal phases of facial expressions and facial action units (AUs) and 2) temporal alignment of a certain facial behavior displayed by two different persons. To this end, the slow features nonnegative matrix factorization (SFNMF) is proposed in order to learn slow varying parts-based representations of time varying sequences capturing the underlying dynamics of temporal phenomena, such as facial expressions. Moreover, the SFNMF is extended in order to handle two temporally misaligned data sequences depicting the same visual phenomena. To do so, the dynamic time warping is incorporated into the SFNMF, allowing the temporal alignment of the data sets onto the subspace spanned by the estimated nonnegative shared latent features amongst the two visual sequences. Extensive experimental results in two video databases demonstrate the effectiveness of the proposed methods in: 1) unsupervised detection of the temporal phases of posed and spontaneous facial events and 2) temporal alignment of facial expressions, outperforming by a large margin the state-of-the-art methods that they are compared to.

Joint modeling of the intensity of facial action units (AUs) from face images is challenging due to the large number of AUs (30+) and their intensity levels (6). This is in part due to the lack of suitable models that can efficiently handle such a large number of outputs/classes simultaneously, but also due to the lack of labelled target data. For this reason, majority of the methods proposed so far resort to independent classifiers for the AU intensity. This is suboptimal for at least two reasons: the facial appearance of some AUs changes depending on the intensity of other AUs, and some AUs co-occur more often than others. Encoding this is expected to improve the estimation of target AU intensities, especially in the case of noisy image features, head-pose variations and imbalanced training data. To this end, we introduce a novel modeling framework, Copula Ordinal Regression (COR), that leverages the power of copula functions and CRFs, to detangle the probabilistic modeling of AU dependencies from the marginal modeling of the AU intensity. Consequently, the COR model achieves the joint learning and inference of intensities of multiple AUs, while being computationally tractable. We show on two challenging datasets of naturalistic facial expressions that the proposed approach consistently outperforms (i) independent modeling of AU intensities, and (ii) the state-of the-art approach for the target task.

Typically, the problems of spatial and temporal alignment of sequences are considered disjoint. That is, in order to align two sequences, a methodology that (non)-rigidly aligns the images is first applied, followed by temporal alignment of the obtained aligned images. In this paper, we propose the first, to the best of our knowledge, methodology that can jointly spatio-temporally align two sequences, which display highly deformable texture-varying objects. We show that by treating the problems of deformable spatial and temporal alignment jointly, we achieve better results than considering the problems independent. Furthermore, we show that deformable spatio-temporal alignment of faces can be performed in an unsupervised manner (i.e., without employing face trackers or building person-specific deformable models).

Deep bottleneck features (DBNFs) have been used successfully in the past for acoustic speech recognition from audio. However, research on extracting DBNFs for visual speech recognition is very limited. In this work, we present an approach to extract deep bottleneck visual features based on deep autoencoders. To the best of our knowledge, this is the first work that extracts DBNFs for visual speech recognition directly from pixels. We first train a deep autoencoder with a bottleneck layer in order to reduce the dimensionality of the image. Then the autoencoder's decoding layers are replaced by classification layers which make the bottleneck features more discriminative. Discrete Cosine Transform (DCT) features are also appended in the bottleneck layer during training in order to make the bottleneck features complementary to DCT features. Long-Short Term Memory (LSTM) networks are used to model the temporal dynamics and the performance is evaluated on the OuluVS and AVLetters databases. The extracted complementary DBNF in combination with DCT features achieve the best performance resulting in an absolute improvement of up to 5% over the DCT baseline.

The ability to automatically detect the extent of agreement or disagreement a person expresses is an important indicator of inter-personal relations and emotion expression. Most of existing methods for automated analysis of human agreement from audio-visual data perform agreement detection using either audio or visual modality of human interactions. However, this is suboptimal as expression of different agreement levels is composed of various facial and vocal cues specific to the target level. To this end, we propose the first approach for multi-modal estimation of agreement intensity levels. Specifically, our model leverages the feature representation power of Multi-modal Neural Networks (NN) and discriminative power of Conditional Ordinal Random Fields (CORF) to achieve dynamic classification of agreement levels from videos. We show on the MAHNOB-Mimicry database of dyadic human interactions that the proposed approach outperforms its uni-modal and linear counterparts, and related models that can be applied to the target task.

Social and Affective Robotics is a growing multidisciplinary field encompassing computer science, engineering, psychology, education, and many other disciplines. It explores how social and affective factors influence interactions between humans and robots, and how affect and social signals can be sensed and integrated into the design, implementation, and evaluation of robots. With talks by renowned researchers in this area, Social and Affective Robotics Tutorial will help both new and experienced researchers to identify trends, concepts, methodologies and applications in this field, identified as a technological megatrend driving the fourth industrial revolution.

The sixth Audio-Visual Emotion Challenge and workshop AVEC 2016 was held in conjunction ACM Multimedia'16. This year the AVEC series addresses two distinct sub-challenges, multi-modal emotion recognition and audio-visual depression detection. Both sub-challenges are in a way a return to AVEC's past editions: the emotion sub-challenge is based on the same dataset as the one used in AVEC 2015, and depression analysis was previously addressed in AVEC 2013/2014. In this summary, we mainly describe participation and its conditions.

Certain inner feelings and physiological states like pain are subjective states that cannot be directly measured, but can be estimated from spontaneous facial expressions. Since they are typically characterized by subtle movements of facial parts, analysis of the facial details is required. To this end, we formulate a new regression method for continuous estimation of the intensity of facial behavior interpretation, called Doubly Sparse Relevance Vector Machine (DSRVM). DSRVM enforces double sparsity by jointly selecting the most relevant training examples (a.k.a. relevance vectors) and the most important kernels associated with facial parts relevant for interpretation of observed facial expressions. This advances prior work on multi-kernel learning, where sparsity of relevant kernels is typically ignored. Empirical evaluation on challenging Shoulder Pain videos, and the benchmark DISFA and SEMAINE datasets demonstrate that DSRVM outperforms competing approaches with a multi-fold reduction of running times in training and testing.

Recovering correlated and individual components of two, possibly temporally misaligned, sets of data is a fundamental task in disciplines such as image, vision, and behavior computing, with application to problems such as multi-modal fusion (via correlated components), predictive analysis, and clustering (via the individual ones). Here, we study the extraction of correlated and individual components under real-world conditions, namely i) the presence of gross non-Gaussian noise and ii) temporally misaligned data. In this light, we propose a method for the Robust Correlated and Individual Component Analysis (RCICA) of two sets of data in the presence of gross, sparse errors. We furthermore extend RCICA in order to handle temporal incongruities arising in the data. To this end, two suitable optimization problems are solved. The generality of the proposed methods is demonstrated by applying them onto 4 applications, namely i) heterogeneous face recognition, ii) multi-modal feature fusion for human behavior analysis (i.e., audio-visual prediction of interest and conflict), iii) face clustering, and iv) thetemporal alignment of facial expressions. Experimental results on 2 synthetic and 7 real world datasets indicate the robustness and effectiveness of the proposed methodson these application domains, outperforming other state-of-the-art methods in the field.

Pain-related emotions are a major barrier to effective self rehabilitation in chronic pain. Automated coaching systems capable of detecting these emotions are a potential solution. This paper lays the foundation for the development of such systems by making three contributions. First, through literature reviews, an overview of how pain is expressed in chronic pain and the motivation for detecting it in physical rehabilitation is provided. Second, a fully labelled multimodal dataset (named `EmoPain') containing high resolution multiple-view face videos, head mounted and room audio signals, full body 3D motion capture and electromyographic signals from back muscles is supplied. Natural unconstrained pain related facial expressions and body movement behaviours were elicited from people with chronic pain carrying out physical exercises. Both instructed and non-instructed exercises were considered to reflect traditional scenarios of physiotherapist directed therapy and home-based self-directed therapy. Two sets of labels were assigned: level of pain from facial expressions annotated by eight raters and the occurrence of six pain-related body behaviours segmented by four experts. Third, through exploratory experiments grounded in the data, the factors and challenges in the automated recognition of such expressions and behaviour are described, the paper concludes by discussing potential avenues in the context of these findings also highlighting differences for the two exercise scenarios addressed.

Prediction plays a key role in recent computational models of the brain and it has been suggested that the brain constantly makes multisensory spatiotemporal predictions. Inspired by these findings we tackle the problem of audiovisual fusion from a new perspective based on prediction. We train predictive models which model the spatiotemporal relationship between audio and visual features by learning the audio-to-visual and visual-to-audio feature mapping for each class. Similarly, we train predictive models which model the time evolution of audio and visual features by learning the past-to-future feature mapping for each class. In classification, all the class-specific regression models produce a prediction of the expected audio/visual features and their prediction errors are combined for each class. The set of class-specific regressors which best describes the audiovisual feature relationship, i.e., results in the lowest prediction error, is chosen to label the input frame. We perform cross-database experiments, using the AMI, SAL, and MAHNOB databases, in order to classify laughter and speech and subject-independent experiments on the AVIC database in order to classify laughter, hesitation and consent. In virtually all cases prediction-based audiovisual fusion consistently outperforms the two most commonly used fusion approaches, decision-level and feature-level fusion.

Emotions are time varying affective phenomena that are elicited as a result of stimuli. Videos and movies in particular are made to elicit emotions in their audiences. Detecting the viewers' emotions instantaneously can be used to find the emotional traces of videos. In this paper, we present our approach in instantaneously detecting the emotions of video viewers' emotions from electroencephalogram (EEG) signals and facial expressions. A set of emotion inducing videos were shown to participants while their facial expressions and physiological responses were recorded. The expressed valence (negative to positive emotions) in the videos of participants' faces were annotated by five annotators. The stimuli videos were also continuously annotated on valence and arousal dimensions. Long-short-term-memory recurrent neural networks (LSTM-RNN) and continuous conditional random fields (CCRF) were utilized in detecting emotions automatically and continuously. We found the results from facial expressions to be superior to the results from EEG signals. We analyzed the effect of the contamination of facial muscle activities on EEG signals and found that most of the emotionally valuable content in EEG features are as a result of this contamination. However, our statistical analysis showed that EEG signals still carry complementary information in presence of facial expressions.

Accent is a soft biometric trait that can be inferred from pronunciation and articulation patterns characterizing the speaking style of an individual. Past research has addressed the task of classifying accent, as belonging to a native language speaker or a foreign language speaker, by means of the audio modality only. However, features extracted from the visual stream of speech have been successfully used to extend or substitute audio-only approaches that target speech or language recognition. Motivated by these findings, we investigate to what extent temporal visual speech dynamics attributed to accent can be modeled and identified when the audio stream is missing or noisy, and the speech content is unknown. We present here a fully automated approach to discriminating native from non-native English speech, based exclusively on visual cues. A systematic evaluation of various appearance and shape features for the target problem is conducted, with the former consistently yielding superior performance. Subject-independent cross-validation experiments are conducted on mobile phone recordings of continuous speech and isolated word utterances spoken by 56 subjects from the challenging MOBIO database. High performance is achieved on a text-dependent (TD) protocol, with the best score of 76.5% yielded by fusion of five hidden Markov models trained on appearance features. Our framework is also efficient even when tested on examples of speech unseen in the training phase, although performing less accurately compared to the TD case.

Automated analysis of facial expressions can benefit many domains, from marketing to clinical diagnosis of neurodevelopmental disorders. Facial expressions are typically encoded as a combination of facial muscle activations, i.e., action units. Depending on context, these action units co-occur in specific patterns, and rarely in isolation. Yet, most existing methods for automatic action unit detection fail to exploit dependencies among them, and the corresponding facial features. To address this, we propose a novel multi-conditional latent variable model for simultaneous fusion of facial features and joint action unit detection. Specifically, the proposed model performs feature fusion in a generative fashion via a low-dimensional shared subspace, while simultaneously performing action unit detection using a discriminative classification approach. We show that by combining the merits of both approaches, the proposed methodology outperforms existing purely discriminative/generative methods for the target task. To reduce the number of parameters, and avoid overfitting, a novel Bayesian learning approach based on Monte Carlo sampling is proposed, to integrate out the shared subspace. We validate the proposed method on posed and spontaneous data from three publicly available datasets (CK+, DISFA and Shoulder-pain), and show that both feature fusion and joint learning of action units leads to improved performance compared to the state-of-the-art methods for the task.

Face images convey rich information which can be perceived as a superposition of low-complexity components associated with attributes, such as facial identity, expressions, and activation of facial action units (AUs). For instance, low-rank components characterizing neutral facial images are associated with identity, while sparse components capturing non-rigid deformations occurring in certain face regions reveal expressions and AU activations. In this paper, the discriminant incoherent component analysis (DICA) is proposed in order to extract low-complexity components, corresponding to facial attributes, which are mutually incoherent among different classes (e.g., identity, expression, and AU activation) from training data, even in the presence of gross sparse errors. To this end, a suitable optimization problem, involving the minimization of nuclear-and l1 -norm, is solved. Having found an ensemble of class-specific incoherent components by the DICA, an unseen (test) image is expressed as a group-sparse linear combination of these components, where the non-zero coefficients reveal the class(es) of the respective facial attribute(s) that it belongs to. The performance of the DICA is experimentally assessed on both synthetic and real-world data. Emphasis is placed on face analysis tasks, namely, joint face and expression recognition, face recognition under varying percentages of training data corruption, subject-independent expression recognition, and AU detection by conducting experiments on four data sets. The proposed method outperforms all the methods that are compared with all the tasks and experimental settings.

A recently introduced latent feature learning technique for time-varying dynamic phenomena analysis is the so-called slow feature analysis (SFA). SFA is a deterministic component analysis technique for multidimensional sequences that, by minimizing the variance of the first-order time derivative approximation of the latent variables, finds uncorrelated projections that extract slowly varying features ordered by their temporal consistency and constancy. In this paper, we propose a number of extensions in both the deterministic and the probabilistic SFA optimization frameworks. In particular, we derive a novel deterministic SFA algorithm that is able to identify linear projections that extract the common slowest varying features of two or more sequences. In addition, we propose an expectation maximization (EM) algorithm to perform inference in a probabilistic formulation of SFA and similarly extend it in order to handle two and more time-varying data sequences. Moreover, we demonstrate that the probabilistic SFA (EM-SFA) algorithm that discovers the common slowest varying latent space of multiple sequences can be combined with dynamic time warping techniques for robust sequence time-alignment. The proposed SFA algorithms were applied for facial behavior analysis, demonstrating their usefulness and appropriateness for this task.

We present a novel approach to automated estimation of agreement intensity levels from facial images. To this end, we employ the MAHNOB Mimicry database of subjects recorded during dyadic interactions, where the facial images are annotated in terms of agreement intensity levels using the Likert scale (strong disagreement, disagreement, neutral, agreement and strong agreement). Dynamic modelling of the agreement levels is accomplished by means of a Conditional Ordinal Random Field model. Specifically, we propose a novel Neural Conditional Ordinal Random Field model that performs non-linear feature extraction from face images using the notion of Neural Networks, while also modelling temporal and ordinal relationships between the agreement levels. We show in our experiments that the proposed approach outperforms existing methods for modelling of sequential data. The preliminary results obtained on five subjects demonstrate that the intensity of agreement can successfully be estimated from facial images (39% F1 score) using the proposed method.

This paper describes a substantial effort to build a real-time interactive multimodal dialogue system with a focus on emotional and non-verbal interaction capabilities. The work is motivated by the aim to provide technology with competences in perceiving and producing the emotional and non-verbal behaviours required to sustain a conversational dialogue. We present the Sensitive Artificial Listener (SAL) scenario as a setting which seems particularly suited for the study of emotional and non-verbal behaviour, since it requires only very limited verbal understanding on the part of the machine. This scenario allows us to concentrate on non-verbal capabilities without having to address at the same time the challenges of spoken language understanding, task modeling etc. We first summarise three prototype versions of the SAL scenario, in which the behaviour of the Sensitive Artificial Listener characters was determined by a human operator. These prototypes served the purpose of verifying the effectiveness of the SAL scenario and allowed us to collect data required for building system components for analysing and synthesising the respective behaviours. We then describe the fully autonomous integrated real-time system we created, which combines incremental analysis of user behaviour, dialogue management, and synthesis of speaker and listener behaviour of a SAL character displayed as a virtual agent. We discuss principles that should underlie the evaluation of SAL-type systems. Since the system is designed for modularity and reuse, and since it is publicly available, the SAL system has potential as a joint research tool in the affective computing research community.

The ability of robots to interact in a socially intelligent manner with humans is the core of human-robot interaction (HRI). The quality of this interaction is typically measured in terms of how it is engaging to the users either reflected in duration of time users spend interacting with a robot, or their self-reports on engagement during the interaction. In contrast to existing studies that analyze the influence of robots' ability to mimic affective states (happy or sad) of users on their engagement, in this paper we study the influence of sentiment apprehension by robots (i.e., robot's ability to reason about the user's attitudes such as judgment / liking) on the user engagement. Specifically, we present the findings from our pilot study on the effect of sentiment apprehension in HRI using NAO robot. In this study, we analyzed two versions of mimicry game: in the first, NAO was solely mimicking facial expressions of the users, while in the second he was also providing a feedback based on the sentiment apprehension. A total of 32 participants (7 female, 25 male) were recruited for this experiment, and the results show that the participants in the second group spent more time interacting with the robot and played more rounds of the mimicry game. After experiencing both versions of the game, ratings given by the participants indicate (with 99% confidence) that the game with sentiment apprehension is more engaging than the baseline version.

We present a user-independent emotion recognition method with the goal of detecting expected emotions or affective tags for videos using electroencephalogram (EEG), pupillary response and gaze distance. We first selected 20 video clips with extrinsic emotional content from movies and online resources. Then EEG responses and eye gaze data were recorded from 24 participants while watching emotional video clips. Ground truth was defined based on the median arousal and valence scores given to clips in a preliminary study. The arousal classes were calm, medium aroused and activated and the valence classes were unpleasant, neutral and pleasant. A one-participant-out cross validation was employed to evaluate the classification performance in a user-independent approach. The best classification accuracy of 68.5% for three labels of valence and 76.4% for three labels of arousal were obtained using a modality fusion strategy and a support vector machine. The results over a population of 24 participants demonstrate that user-independent emotion recognition can outperform individual self-reports for arousal assessments and do not underperform for valence assessments.

This paper is about estimating intensity levels of Facial Action Units (FAUs) in videos as an important step toward interpreting facial expressions. As input features, we use locations of facial landmark points detected in video frames. To address uncertainty of input, we formulate a generative latent tree (LT) model, its inference, and novel algorithms for efficient learning of both LT parameters and structure. Our structure learning iteratively builds LT by adding either a new edge or a new hidden node to LT, starting from initially independent nodes of observable features. A graph-edit operation that increases maximally the likelihood and minimally the model complexity is selected as optimal in each iteration. For FAU intensity estimation, we derive closed-form expressions of posterior marginals of all variables in LT, and specify an efficient bottom-up/top-down inference. Our evaluation on the benchmark DISFA and ShoulderPain datasets, in subject-independent setting, demonstrate that we outperform the state of the art, even under significant noise in facial landmarks. Effectiveness of our structure learning is demonstrated by probabilistically sampling meaningful facial expressions from the LT.

We propose a novel multi-conditional latent variable model for simultaneous facial feature fusion and detection of facial action units. In our approach we exploit the structure-discovery capabilities of generative models such as Gaussian processes, and the discriminative power of classifiers such as logistic function. This leads to superior performance compared to existing classifiers for the target task that exploit either the discriminative or generative property, but not both. The model learning is performed via an efficient, newly proposed Bayesian learning strategy based on Monte Carlo sampling. Consequently, the learned model is robust to data overfitting, regardless of the number of both input features and jointly estimated facial action units. Extensive qualitative and quantitative experimental evaluations are performed on three publicly available datasets (CK+, Shoulder-pain and DISFA). We show that the proposed model outperforms the state-of-the-art methods for the target task on (i) feature fusion, and (ii) multiple facial action unit detection.

Recently, it has been shown that excellent results can be achieved in both facial landmark localization and pose-invariant face recognition. These breakthroughs are attributed to the efforts of the community to manually annotate facial images in many different poses and to collect 3D facial data. In this paper, we propose a novel method for joint frontal view reconstruction and landmark localization using a small set of frontal images only. By observing that the frontal facial image is the one having the minimum rank of all different poses, an appropriate model which is able to jointly recover the frontalized version of the face as well as the facial landmarks is devised. To this end, a suitable optimization problem, involving the minimization of the nuclear norm and the matrix l1 norm is solved. The proposed method is assessed in frontal face reconstruction, face landmark localization, pose-invariant face recognition, and face verification in unconstrained conditions. The relevant experiments have been conducted on 8 databases. The experimental results demonstrate the effectiveness of the proposed method in comparison to the state-of-the-art methods for the target problems.

Finding landmarks on objects like faces is a challenging computer vision problem, especially in real life conditions (or in-the-wild) and Active Appearance Models have been widely used to solve it. State-of-the-art algorithms for fitting an AAM to a new image are based on Gauss-Newton (GN) optimization. Recently fast GN algorithms have been proposed for both forward additive and inverse compositional fitting frameworks. In this paper, we propose a fast and exact bi-directional (Fast-Bd) approach to AAM fitting by combining both approaches. Although such a method might appear to increase computational burden, we show that by capitalizing on results from optimization theory, an exact solution, as computationally efficient as the original forward or inverse formulation, can be derived. Our proposed bi-directional approach achieves state-of-the-art performance and superior convergence properties. These findings are validated on two challenging, in-the-wild data sets, LFPW and Helen, and comparison is provided to the state-of-the art methods for Active Appearance Models fitting.

The fifth Audio-Visual Emotion Challenge and workshop AVEC 2015 was held in conjunction ACM Multimedia'15. Like the previous editions of AVEC, the workshop/challenge addresses the detection of affective signals represented in audio-visual data in terms of high-level continuous dimensions. A major novelty was further introduced this year by the inclusion of the physiological modality - along with the audio and the video modalities - in the dataset. In this summary, we mainly describe participation and its conditions.

We propose a face alignment framework that relies on the texture model generated by the responses of discriminatively trained part-based filters. Unlike standard texture models built from pixel intensities or responses generated by generic filters (e.g. Gabor), our framework has two important advantages. First, by virtue of discriminative training, invariance to external variations (like identity, pose, illumination and expression) is achieved. Second, we show that the responses generated by discriminatively trained filters (or patch-experts) are sparse and can be modeled using a very small number of parameters. As a result, the optimization methods based on the proposed texture model can better cope with unseen variations. We illustrate this point by formulating both part-based and holistic approaches for generic face alignment and show that our framework outperforms the state-of-the-art on multiple "wild" databases. The code and dataset annotations are available for research purposes from http://ibug.doc.ic.ac.uk/resources.

Hidden conditional random fields (HCRFs) are discriminative latent variable models which have been shown to successfully learn the hidden structure of a given classification problem. An Infinite hidden conditional random field is a hidden conditional random field with a countably infinite number of hidden states, which rids us not only of the necessity to specify a priori a fixed number of hidden states available but also of the problem of overfitting. Markov chain Monte Carlo (MCMC) sampling algorithms are often employed for inference in such models. However, convergence of such algorithms is rather difficult to verify, and as the complexity of the task at hand increases the computational cost of such algorithms often becomes prohibitive. These limitations can be overcome by variational techniques. In this paper, we present a generalized framework for infinite HCRF models, and a novel variational inference approach on a model based on coupled Dirichlet Process Mixtures, the HCRF-DPM. We show that the variational HCRF-DPM is able to converge to a correct number of represented hidden states, and performs as well as the best parametric HCRFs-chosen via cross-validation-for the difficult tasks of recognizing instances of agreement, disagreement, and pain in audiovisual sequences.

Modeling intensity of facial action units from spontaneously displayed facial expressions is challenging mainly because of high variability in subject-specific facial expressiveness, head-movements, illumination changes, etc. These factors make the target problem highly context-sensitive. However, existing methods usually ignore this context-sensitivity of the target problem. We propose a novel Conditional Ordinal Random Field (CORF) model for context-sensitive modeling of the facial action unit intensity, where the W5+ (who, when, what, where, why and how) definition of the context is used. While the proposed model is general enough to handle all six context questions, in this paper we focus on the context questions: who (the observed subject), how (the changes in facial expressions), and when (the timing of facial expressions and their intensity). The context questions who and howare modeled by means of the newly introduced context-dependent covariate effects, and the context question when is modeled in terms of temporal correlation between the ordinal outputs, i.e., intensity levels of action units. We also introduce a weighted softmax-margin learning of CRFs from data with skewed distribution of the intensity levels, which is commonly encountered in spontaneous facial data. The proposed model is evaluated on intensity estimation of pain and facial action units using two recently published datasets (UNBC Shoulder Pain and DISFA) of spontaneously displayed facial expressions. Our experiments show that the proposed model performs significantly better on the target tasks compared to the state-of-the-art approaches. Furthermore, compared to traditional learning of CRFs, we show that the proposed weighted learning results in more robust parameter estimation from the imbalanced intensity data.

The papers in this special section were presented at the 6th Biannual International Conference on Affective Computing and Intelligent Interaction (ACII 2013), held from September 2-5, 2013, in Geneva, Switzerland. These papers reflect the diversity of affective computing research, ranging from physiological models for affect detection, methodological issues, computational models of affect, and human perceptions of affective virtual agents.

Images of facial expressions are often captured from various views as a result of either head movements or variable camera position. Existing methods for multiview and/or view-invariant facial expression recognition typically perform classification of the observed expression using either classifiers learned separately for each view or a single classifier learned for all views. However, these approaches ignore the fact that different views of a facial expression are just different manifestations of the same facial expression. By accounting for this redundancy, we can design more effective classifiers for the target task. To this end, we propose a discriminative shared Gaussian process latent variable model (DS-GPLVM) for multiview and view-invariant classification of facial expressions from multiple views. In this model, we first learn a discriminative manifold shared by multiple views of a facial expression. Subsequently, we perform facial expression classification in the expression manifold. Finally, classification of an observed facial expression is carried out either in the view-invariant manner (using only a single view of the expression) or in the multiview manner (using multiple views of the expression). The proposed model can also be used to perform fusion of different facial features in a principled manner. We validate the proposed DS-GPLVM on both posed and spontaneously displayed facial expressions from three publicly available datasets (MultiPIE, labeled face parts in the wild, and static facial expressions in the wild). We show that this model outperforms the state-of-the-art methods for multiview and view-invariant facial expression classification, and several state-of-the-art methods for multiview learning and feature fusion.

Slow feature analysis (SFA) is a dimensionality reduction technique which has been linked to how visual brain cells work. In recent years, the SFA was adopted for computer vision tasks. In this paper, we propose an exact kernel SFA (KSFA) framework for positive definite and indefinite kernels in Krein space. We then formulate an online KSFA which employs a reduced set expansion. Finally, by utilizing a special kind of kernel family, we formulate exact online KSFA for which no reduced set is required. We apply the proposed system to develop a SFA-based change detection algorithm for stream data. This framework is employed for temporal video segmentation and tracking. We test our setup on synthetic and real data streams. When combined with an online learning tracking system, the proposed change detection approach improves upon tracking setups that do not utilize change detection.

The development of facial databases with an abundance of annotated facial data captured under unconstrained 'in-the-wild' conditions have made discriminative facial deformable models the de facto choice for generic facial landmark localization. Even though very good performance for the facial landmark localization has been shown by many recently proposed discriminative techniques, when it comes to the applications that require excellent accuracy, such as facial behaviour analysis and facial motion capture, the semi-automatic person-specific or even tedious manual tracking is still the preferred choice. One way to construct a person-specific model automatically is through incremental updating of the generic model. This paper deals with the problem of updating a discriminative facial deformable model, a problem that has not been thoroughly studied in the literature. In particular, we study for the first time, to the best of our knowledge, the strategies to update a discriminative model that is trained by a cascade of regressors. We propose very efficient strategies to update the model and we show that is possible to automatically construct robust discriminative person and imaging condition specific models 'in-the-wild' that outperform state-of-the-art generic face alignment strategies.

In this paper we introduce a new distance for robustly matching vectors of 3D rotations. A special representation of 3D rotations, which we coin full-angle quaternion (FAQ), allows us to express this distance as Euclidean. We apply the distance to the problems of 3D shape recognition from point clouds and 2D object tracking in color video. For the former, we introduce a hashing scheme for scale and translation which outperforms the previous state-of-the-art approach on a public dataset. For the latter, we incorporate online subspace learning with the proposed FAQ representation to highlight the benefits of the new representation.

A key problem often encountered by many learning algorithms in computer vision dealing with high dimensional data is the so called "curse of dimensionality" which arises when the available training samples are less than the input feature space dimensionality. To remedy this problem, we propose a joint dimensionality reduction and classification framework by formulating an optimization problem within the maximum margin class separation task. The proposed optimization problem is solved using alternative optimization where we jointly compute the low dimensional maximum margin projections and the separating hyperplanes in the projection subspace. Moreover, in order to reduce the computational cost of the developed optimization algorithm we incorporate orthogonality constraints on the derived projection bases and show that the resulting combined model is an alternation between identifying the optimal separating hyperplanes and performing a linear discriminant analysis on the support vectors. Experiments on face, facial expression and object recognition validate the effectiveness of the proposed method against state-of-the-art dimensionality reduction algorithms.

The construction of Facial Deformable Models (FDMs) is a very challenging computer vision problem, since the face is a highly deformable object and its appearance drastically changes under different poses, expressions, and illuminations. Although several methods for generic FDMs construction, have been proposed for facial landmark localization in still images, they are insufficient for tasks such as facial behaviour analysis and facial motion capture where perfect landmark localization is required. In this case, person-specific FDMs (PSMs) are mainly employed, requiring manual facial landmark annotation for each person and person-specific training. In this paper, a novel method for the automatic construction of PSMs is proposed. To this end, an orthonormal subspace which is suitable for facial image reconstruction is learnt. Next, to correct the fittings of a generic model, image congealing (i.e., batch image aliment) is performed by employing only the learnt orthonormal subspace. Finally, the corrected fittings are used to construct the PSM. The image congealing problem is solved by formulating a suitable sparsity regularized rank minimization problem. The proposed method outperforms the state-of-the art methods that is compared to, in terms of both landmark localization accuracy and computational time.

Arguably, Deformable Part Models (DPMs) are one of the most prominent approaches for face alignment with impressive results being recently reported for both controlled lab and unconstrained settings. Fitting in most DPM methods is typically formulated as a two-step process during which discriminatively trained part templates are first correlated with the image to yield a filter response for each landmark and then shape optimization is performed over these filter responses. This process, although computationally efficient, is based on fixed part templates which are assumed to be independent, and has been shown to result in imperfect filter responses and detection ambiguities. To address this limitation, in this paper, we propose to jointly optimize a part-based, trained in-the-wild, flexible appearance model along with a global shape model which results in a joint translational motion model for the model parts via Gauss-Newton (GN) optimization. We show how significant computational reductions can be achieved by building a full model during training but then efficiently optimizing the proposed cost function on a sparse grid using weighted least-squares during fitting. We coin the proposed formulation Gauss-Newton Deformable Part Model (GN-DPM). Finally, we compare its performance against the state-of-the-art and show that the proposed GN-DPM outperforms it, in some cases, by a large margin. Code for our method is available from http://ibug.doc.ic.ac.uk/resources.

This video details the development of an intelligent outdoor Guide robot. The main objective is to deploy an innovative robotic guide which is not only able to show information, but to react to the affective states of the users, and to offer location-based services using augmented reality. The scientific challenges concern autonomous outdoor navigation and localization, robust 24/7 operation, affective interaction with visitors through outdoor human and facial feature detection as well as engaging interactive behaviors in an ongoing non-verbal dialogue with the user.

Facial behaviour is our preeminent means to communicating affective and social signals. This talk discusses a number of components of human facial behaviour, how they can be automatically sensed and analysed by computers, what is the past research in the field conducted by the iBUG group at Imperial College London, and how far we are from enabling computers to sense and recognise human facial expressions and behaviour.

Accent is an important biometric characteristic that is defined by the presence of specific traits in the speaking style of an individual. These are identified by patterns in the speech production system, such as those present in the vocal tract or in lip movements. Evidence from linguistics and speech processing research suggests that visual information enhances speech recognition. Intrigued by these findings, along with the assumption that visually perceivable accent-related patterns are transferred from the mother tongue to a foreign language, we investigate the task of discriminating native from non-native speech in English, employing visual features only. Training and evaluation is performed on segments of continuous visual speech, captured by mobile phones, where all speakers read the same text. We apply various appearance descriptors to represent the mouth region at each video frame. Vocabulary-based histograms, being the final representation of dynamic features for all utterances, are used for recognition. Binary classification experiments, discriminating native and non-native speakers, are conducted in a subject-independent manner. Our results show that this task can be addressed by means of an automated approach that uses visual features only.

The problem of automatically estimating the interest level of a subject has been gaining attention by researchers, mostly due to the vast applicability of interest detection. In this work, we obtain a set of continuous interest annotations for the SE-MAINE database, which we analyse also in terms of emotion dimensions such as valence and arousal. Most importantly, we propose a robust variant of Canonical Correlation Analysis (RCCA) for performing audio-visual fusion, which we apply to the prediction of interest. RCCA recovers a low-rank subspace which captures the correlations of fused modalities, while isolating gross errors in the data without making any assumptions regarding Gaussianity. We experimentally show that RCCA is more appropriate than other standard fusion techniques (such as l2-CCA and feature-level fusion), since it both captures interactions between modalities while also decontaminating the obtained subspace from errors which are dominant in real-world problems.

We propose a 3D Constrained Local Model framework for deformable face alignment in depth image. Our framework exploits the intrinsic 3D geometric information in depth data by utilizing robust histogram-based 3D geometric features that are based on normal vectors. In addition, we demonstrate the fusion of intensity data and 3D features that further improves the facial landmark localization accuracy. The experiments are conducted on publicly available FRGC database. The results show that our 3D features based CLM completely outperforms the raw depth features based CLM in term of fitting accuracy and robustness, and the fusion of intensity and 3D depth feature further improves the performance. Another benefit is that the proposed 3D features in our framework do not require any pre-processing procedure on the data.

Active Appearance Models (AAMs) are statistical models of shape and appearance widely used in computer vision to detect landmarks on objects like faces. Fitting an AAM to a new image can be formulated as a non-linear least-squares problem which is typically solved using iterative methods. Owing to its efficiency, Gauss-Newton optimization has been the standard choice over more sophisticated approaches like Newton. In this paper, we show that the AAM problem has structure which can be used to solve efficiently the original Newton problem without any approximations. We then make connections to the original Gauss-Newton algorithm and study experimentally the effect of the additional terms introduced by the Newton formulation on both fitting accuracy and convergence. Based on our derivations, we also propose a combined Newton and Gauss-Newton method which achieves promising fitting and convergence performance. Our findings are validated on two challenging in-the-wild data sets.

In this paper, we combine the principles of temporal slowness and nonnegative parts-based learning into a single framework that aims to learn slow varying parts-based representations of time varying sequences. We demonstrate that the proposed algorithm arises naturally by embedding the Slow Features Analysis trace optimization problem in the nonnegative subspace learning framework and derive novel multiplicative update rules for its optimization. The usefulness of the developed algorithm is demonstrated for unsupervised facial behaviour dynamics analysis on MMI database.

Emotions play an important role in how we select and consume multimedia. Recent advances on affect detection are focused on detecting emotions continuously. In this paper, for the first time, we continuously detect valence from electroencephalogram (EEG) signals and facial expressions in response to videos. Multiple annotators provided valence levels continuously by watching the frontal facial videos of participants who watched short emotional videos. Power spectral features from EEG signals as well as facial fiducial points are used as features to detect valence levels for each frame continuously. We study the correlation between features from EEG and facial expressions with continuous valence. We have also verified our model's performance for the emotional highlight detection using emotion recognition from EEG signals. Finally the results of multimodal fusion between facial expression and EEG signals are presented. Having such models we will be able to detect spontaneous and subtle affective responses over time and use them for video highlight detection.

In this paper we propose a new method for the detection of action units that relies on a novel region-based face representation and a mid-level decision layer that combines region-specific information. Different from other approaches, we do not represent the face as a regular grid based on the face location alone (holistic representation), nor by using small patches centred at iducial facial point locations (local representation). Instead, we propose to use domain knowledge regarding AU-specific facial muscle contractions to define a set of face regions covering the whole face. Therefore, as opposed to local appearance models, our face representation makes use of the full facial appearance, while the use of facial point locations to define the regions means that we obtain better-registered descriptors compared to holistic representations. Finally, we propose an AU-specific weighted sum model is used as a decision-level fusion layer in charge of combining region-specific probabilistic information. This configuration allows each classier to learning the typical appearance changes for a specific face part and reduces the dimensionality of the problem thus proving to be more robust. Our approach is evaluated on the DISFA and GEMEP-FERA datasets using two histogram-based appearance features, Local Binary Pattern and Local Phase Quantisation. We show superior performance for both the domain-specific region definition and the decision-level fusion respect to the standard approaches when it comes to automatic facial action unit detection.

The task of classifying accent, as belonging to a native language speaker or a foreign language speaker, has been so far addressed by means of the audio modality only. However, features extracted from the visual modality have been successfully used to extend or substitute audio-only approaches developed for speech or language recognition. This paper presents a fully automated approach to discriminating native from non-native speech in English, based exclusively on visual appearance features from speech. Long Short-Term Memory Neural Networks (LSTMs) are employed to model accent-related speech dynamics and yield accent-class predictions. Subject-independent experiments are conducted on speech episodes captured by mobile phones from the challenging MOBIO Database. We establish a text-dependent scenario, using only those recordings in which all subjects read the same paragraph. Our results show that decision level fusion of networks trained with complementary appearance descriptors consistently leads to performance improvement over single-feature systems, with the highest gain in accuracy reaching 7.3%. The best feature combinations achieve classification accuracy of 75%, rendering the proposed method a useful accent classification tool in cases of missing or noisy audio stream.

The fourth Audio-Visual Emotion Challenge and workshop AVEC 2014 was held in conjunction ACM Multimedia'14. Like the 2013 edition of AVEC, the workshop/challenge addresses the interpretation of social signals represented in both audio and video in terms of high-level continuous dimensions from a large number of clinically depressed patients and controls, with a sub-challenge in self-reported severity of depression estimation. In this summary, we mainly describe participation and its conditions.

We present a robust real-time face tracking system based on the Constrained Local Models framework by adopting the novel regression-based Discriminative Response Map Fitting (DRMF) method. By exploiting the algorithm's potential parallelism, we present a hybrid CPU-GPU implementation capable of achieving real-time performance at 30 to 45 FPS, on ordinary consumer-grade computers. We have made the software publicly available for research purposes

Fusing multiple continuous expert annotations is a crucial problem in machine learning and computer vision, particularly when dealing with uncertain and subjective tasks related to affective behavior. Inspired by the concept of inferring shared and individual latent spaces in Probabilistic Canonical Correlation Analysis (PCCA), we propose a novel, generative model that discovers temporal dependencies on the shared/individual spaces (Dynamic Probabilistic CCA, DPCCA). In order to accommodate for temporal lags, which are prominent amongst continuous annotations, we further introduce a latent warping process, leading to the DPCCA with Time Warpings (DPCTW) model. Finally, we propose two supervised variants of DPCCA/DPCTW which incorporate inputs (i.e., visual or audio features), both in a generative (SG-DPCCA) and discriminative manner (SD-DPCCA). We show that the resulting family of models (i) can be used as a unifying framework for solving the problems of temporal alignment and fusion of multiple annotations in time, (ii) can automatically rank and filter annotations based on latent posteriors or other model statistics, and (iii) that by incorporating dynamics, modeling annotation-specific biases, noise estimation, time warping and supervision, DPCTW outperforms state-of-the-art methods for both the aggregation of multiple, yet imperfect expert annotations as well as the alignment of affective behavior.

Both the configuration and the dynamics of facial expressions are crucial for the interpretation of human facial behavior. Yet to date, the vast majority of reported efforts in the field either do not take the dynamics of facial expressions into account, or focus only on prototypic facial expressions of six basic emotions. Facial dynamics can be explicitly analyzed by detecting the constituent temporal segments in Facial Action Coding System (FACS) Action Units (AUs)-onset, apex, and offset. In this paper, we present a novel approach to explicit analysis of temporal dynamics of facial actions using the dynamic appearance descriptor Local Phase Quantization from Three Orthogonal Planes (LPQ-TOP). Temporal segments are detected by combining a discriminative classifier for detecting the temporal segments on a frame-by-frame basis with Markov Models that enforce temporal consistency over the whole episode. The system is evaluated in detail over the MMI facial expression database, the UNBC-McMaster pain database, the SAL database, the GEMEP-FERA dataset in database-dependent experiments, in cross-database experiments using the Cohn-Kanade, and the SEMAINE databases. The comparison with other state-of-the-art methods shows that the proposed LPQ-TOP method outperforms the other approaches for the problem of AU temporal segment detection, and that overall AU activation detection benefits from dynamic appearance information.

We present Active Orientation Models (AOMs), generative models of facial shape and appearance, which extend the well-known paradigm of Active Appearance Models (AAMs) for the case of generic face alignment under unconstrained conditions. Robustness stems from the fact that the proposed AOMs employ a statistically robust appearance model based on the principal components of image gradient orientations. We show that when incorporated within standard optimization frameworks for AAM learning and fitting, this kernel Principal Component Analysis results in robust algorithms for model fitting. At the same time, the resulting optimization problems maintain the same computational cost. As a result, the main similarity of AOMs with AAMs is the computational complexity. In particular, the project-out version of AOMs is as computationally efficient as the standard project-out inverse compositional algorithm, which is admittedly one of the fastest algorithms for fitting AAMs. We verify experimentally that: 1) AOMs generalize well to unseen variations and 2) outperform all other state-of-the-art AAM methods considered by a large margin. This performance improvement brings AOMs at least in par with other contemporary methods for face alignment. Finally, we provide MATLAB code at http://ibug.doc.ic.ac.uk/resources.

Human mimicry is a behavioural cue occurring during social interaction that can inform us about the participants' inter-personal states and attitudes. It occurs when a participant in an interaction exhibits some behaviour as a result of a co-participants prior display of that signal, and occurs on both short and long time-scales. To develop a detection method for such behaviour, we use a method based on feature prediction, where we train an ensemble of regression models from one subject's features to the co-subject's features, for each class. The ensemble of models with lowest reconstruction error is used to detect mimicry and non-mimicry, using continuous audiovisual streams. As mimicry events are dynamical phenomena, we use a temporal regression model (long short-term memory neural networks) to capture sequential dependencies in the data. On a data set of ten 12-minute dyadic interaction episodes, our method gave average positive and negative recall rates of 77.5% and 60.0% respectively, on data with significant class imbalances, due to the relative sparsity of mimicry samples when doing continuous detection.

Laughter is clearly an audiovisual event, consisting of the laughter vocalization and of facial activity, mainly around the mouth and sometimes in the upper face. However, past research on laughter recognition has mainly focused on the information available in the audio channel only, mainly due to the lack of suitable audiovisual data. Only recently few works have been published which combine audio and visual information and most of them deal with the problem of discriminating laughter from speech or other nonlinguistic vocalisations using presegmented data. There are very few works on audiovisual laughter detection from unsegmented audiovisual streams and have either been tested on small datasets or use coarse visual features. As a consequence, results are mixed and it is not clear to what extent the addition of visual information to audio is beneficial for laughter detection. In this work, we attempt to overcome the limitation of previous studies and investigate the performance of audiovisual fusion for laughter detection using audiovisual continuous streams from the SEMAINE database. Our results suggest that there is indeed an improvement in laughter detection with the addition of visual information which is dependent on the performance of the voice activity detector.

We present a novel discriminative regression based approach for the Constrained Local Models (CLMs) framework, referred to as the Discriminative Response Map Fitting (DRMF) method, which shows impressive performance in the generic face fitting scenario. The motivation behind this approach is that, unlike the holistic texture based features used in the discriminative AAM approaches, the response map can be represented by a small set of parameters and these parameters can be very efficiently used for reconstructing unseen response maps. Furthermore, we show that by adopting very simple off-the-shelf regression techniques, it is possible to learn robust functions from response maps to the shape parameters updates. The experiments, conducted on Multi-PIE, XM2VTS and LFPW database, show that the proposed DRMF method outperforms state-of-the-art algorithms for the task of generic face fitting. Moreover, the DRMF method is computationally very efficient and is real-time capable. The current MATLAB implementation takes 1 second per image. To facilitate future comparisons, we release the MATLAB code and the pre-trained models for research purposes.

Temporal alignment of human behaviour from visual data is a very challenging problem due to a numerous reasons, including possible large temporal scale differences, inter/intra subject variability and, more importantly, due to the presence of gross errors and outliers. Gross errors are often in abundance due to incorrect localization and tracking, presence of partial occlusion etc. Furthermore, such errors rarely follow a Gaussian distribution, which is the de-facto assumption in machine learning methods. In this paper, building on recent advances on rank minimization and compressive sensing, a novel, robust to gross errors temporal alignment method is proposed. While previous approaches combine the dynamic time warping (DTW) with low-dimensional projections that maximally correlate two sequences, we aim to learn two underlying projection matrices (one for each sequence), which not only maximally correlate the sequences but, at the same time, efficiently remove the possible corruptions in any datum in the sequences. The projections are obtained by minimizing the weighted sum of nuclear and ℓ1norms, by solving a sequence of convex optimization problems, while the temporal alignment is found by applying the DTW in an alternating fashion. The superiority of the proposed method against the state-of-the-art time alignment methods, namely the canonical time warping and the generalized time warping, is indicated by the experimental results on both synthetic and real datasets.

We describe a very simple framework for deriving the most-well known optimization problems in Active Appearance Models (AAMs), and most importantly for providing efficient solutions. Our formulation results in two optimization problems for fast and exact AAM fitting, and one new algorithm which has the important advantage of being applicable to 3D. We show that the dominant cost for both forward and inverse algorithms is a few times mN which is the cost of projecting an image onto the appearance subspace. This makes both algorithms not only computationally realizable but also very attractive speed-wise for most current systems. Because exact AAM fitting is no longer computationally prohibitive, we trained AAMs in-the-wild with the goal of investigating whether AAMs benefit from such a training process. Our results show that although we did not use sophisticated shape priors, robust features or robust norms for improving performance, AAMs perform notably well and in some cases comparably with current state-of-the-art methods. We provide Matlab source code for training, fitting and reproducing the results presented in this paper at http://ibug.doc.ic.ac.uk/resources.

A recently introduced latent feature learning technique for time varying dynamic phenomena analysis is the so called Slow Feature Analysis (SFA). SFA is a deterministic component analysis technique for multi-dimensional sequences that by minimizing the variance of the first order time derivative approximation of the input signal finds uncorrelated projections that extract slowly-varying features ordered by their temporal consistency and constancy. In this paper, we propose a number of extensions in both the deterministic and the probabilistic SFA optimization frameworks. In particular, we derive a novel deterministic SFA algorithm that is able to identify linear projections that extract the common slowest varying features of two or more sequences. In addition, we propose an Expectation Maximization (EM) algorithm to perform inference in a probabilistic formulation of SFA and similarly extend it in order to handle two and more time varying data sequences. Moreover, we demonstrate that the probabilistic SFA (EMSFA) algorithm that discovers the common slowest varying latent space of multiple sequences can be combined with dynamic time warping techniques for robust sequence time alignment. The proposed SFA algorithms were applied for facial behavior analysis demonstrating their usefulness and appropriateness for this task.

Electroencephalogram (EEG) signals reflect brain activities associated with emotional and cognitive processes. In this paper, we demonstrate how they can be used to find tags for multimedia content without users' direct input. Alternative methods for multimedia tagging is attracting increasing interest from multimedia community. The new portable EEG helmets are paving the way for employing brain waves in human computer interaction. In this paper, we demonstrate the performance of EEG for tagging purposes using two different scenarios on MAHNOB-HCI database. First, an emotional tagging and classification using a reduced set of electrodes is presented. The emotional responses of 24 participants to short video clips are classified into three classes on arousal and valence. We show how a reduced set of electrodes based on previous studies can preserve and even enhance the emotional classification rate. We then demonstrate the feasibility of using EEG signals for tag relevance tasks. A set of images were shown to participants first, without any tag and then with a relevant or irrelevant tag. The relevance of the tag was assessed based on the EEG responses of the participants in the first second after the tag was depicted. Finally, we demonstrate that by aggregating multiple participants' responses we can significantly improve the tagging accuracy.

Adopting continuous dimensional annotations for affective analysis has been gaining rising attention by researchers over the past years. Due to the idiosyncratic nature of this problem, many subproblems have been identified, spanning from the fusion of multiple continuous annotations to exploiting output-correlations amongst emotion dimensions. In this paper, we firstly empirically answer several important questions which have found partial or no answer at all so far in related literature. In more detail, we study the correlation of each emotion dimension (i) with respect to other emotion dimensions, (ii) to basic emotions (e.g., happiness, anger). As a measure for comparison, we use video and audio features. Interestingly enough, we find that (i) each emotion dimension is more correlated with other emotion dimensions rather than with face and audio features, and similarly (ii) that each basic emotion is more correlated with emotion dimensions than with audio and video features. A similar conclusion holds for discrete emotions which are found to be highly correlated to emotion dimensions as compared to audio and/or video features. Motivated by these findings, we present a novel regression algorithm (Correlated-Spaces Regression, CSR), inspired by Canonical Correlation Analysis (CCA) which learns output-correlations and performs supervised dimensionality reduction and multimodal fusion by (i) projecting features extracted from all modalities and labels onto a common space where their inter-correlation is maximised and (ii) learning mappings from the projected feature space onto the projected, uncorrelated label space.

One of the most commonly used audiovisual fusion approaches is feature-level fusion where the audio and visual features are concatenated. Although this approach has been successfully used in several applications, it does not take into account interactions between the features, which can be a problem when one and/or both modalities have noisy features. In this paper, we investigate whether feature fusion based on explicit modelling of interactions between audio and visual features can enhance the performance of the classifier that performs feature fusion using simple concatenation of the audio-visual features. To this end, we propose a log-linear model, named Bimodal Log-linear regression, which accounts for interactions between the features of the two modalities. The performance of the target classifiers is measured in the task of laughter-vs-speech discrimination, since both laughter and speech are naturally audiovisual events. Our experiments on the MAHNOB laughter database suggest that feature fusion based on explicit modelling of interactions between the audio-visual features leads to an improvement of 3\% over the standard feature concatenation approach, when log-linear model is used as the base classifier. Finally, the most and least influential features can be easily identified by observing their interactions.

Users react differently to non-relevant and relevant tags associated with content. These spontaneous reactions can be used for labeling large multimedia databases. We present a method to assess tag relevance to images using the non-verbal bodily responses, namely, electroencephalogram (EEG), facial expressions, and eye gaze. We conducted experiments in which 28 images were shown to 28 subjects once with correct and another time with incorrect tags. The goal of our system is to detect the responses to non-relevant tags and consequently filter them out. Therefore, we trained classifiers to detect the tag relevance from bodily responses. We evaluated the performance of our system using a subject independent approach. The precision at top 5% and top 10% detections were calculated and results of different modalities and different classifiers were compared. The results show that eye gaze outperforms the other modalities in tag relevance detection both overall and for top ranked results.

The third Audio-Visual Emotion Challenge and workshop AVEC 2013 will be held in conjunction ACM Multimedia'13. Like the 2012 edition of AVEC, the workshop/challenge addresses the interpretation of social signals represented in both audio and video in terms of the high-level continuous dimensions arousal and valence, but importantly this year the data is that of a large number of clinically depressed patients and controls, with a sub-challenge in self-reported severity of depression estimation. Like both previous AVECs, the aim is to bring together the audio and video analysis communities.

Principal Component Analysis (PCA) is perhaps the most prominent learning tool for dimensionality reduction in pattern recognition and computer vision. However, the ℓ 2-norm employed by standard PCA is not robust to outliers. In this paper, we propose a kernel PCA method for fast and robust PCA, which we call Euler-PCA (e-PCA). In particular, our algorithm utilizes a robust dissimilarity measure based on the Euler representation of complex numbers. We show that Euler-PCA retains PCA’s desirable properties while suppressing outliers. Moreover, we formulate Euler-PCA in an incremental learning framework which allows for efficient computation. In our experiments we apply Euler-PCA to three different computer vision applications for which our method performs comparably with other state-of-the-art approaches.

We propose a new algorithm to detect facial points in frontal and near-frontal face images. It combines a regression-based approach with a probabilistic graphical model-based face shape model that restricts the search to anthropomorphically consistent regions. While most regression-based approaches perform a sequential approximation of the target location, our algorithm detects the target location by aggregating the estimates obtained from stochastically selected local appearance information into a single robust prediction. The underlying assumption is that by aggregating the different estimates, their errors will cancel out as long as the regressor inputs are uncorrelated. Once this new perspective is adopted, the problem is reformulated as how to optimally select the test locations over which the regressors are evaluated. We propose to extend the regression-based model to provide a quality measure of each prediction, and use the shape model to restrict and correct the sampling region. Our approach combines the low computational cost typical of regression-based approaches with the robustness of exhaustive-search approaches. The proposed algorithm was tested on over 7,500 images from five databases. Results showed significant improvement over the current state of the art.

We propose a method for head-pose invariant facial expression recognition that is based on a set of characteristic facial points. To achieve head-pose invariance, we propose the Coupled Scaled Gaussian Process Regression (CSGPR) model for head-pose normalization. In this model, we first learn independently the mappings between the facial points in each pair of (discrete) nonfrontal poses and the frontal pose, and then perform their coupling in order to capture dependences between them. During inference, the outputs of the coupled functions from different poses are combined using a gating function, devised based on the head-pose estimation for the query points. The proposed model outperforms state-of-the-art regression-based approaches to head-pose normalization, 2D and 3D Point Distribution Models (PDMs), and Active Appearance Models (AAMs), especially in cases of unknown poses and imbalanced training data. To the best of our knowledge, the proposed method is the first one that is able to deal with expressive faces in the range from -45° to +45° pan rotation and -30° to +30° tilt rotation, and with continuous changes in head pose, despite the fact that training was conducted on a small set of discrete poses. We evaluate the proposed method on synthetic and real images depicting acted and spontaneously displayed facial expressions.

This paper presents a novel software framework for the development and research in the area of multimodal human-computer interface (MHCI) systems. The proposed software framework, which is called the HCI∧2 Framework, is built upon publish/subscribe (P/S) architecture. It implements a shared-memory-based data transport protocol for message delivery and a TCP-based system management protocol. The latter ensures that the integrity of system structure is maintained at runtime. With the inclusion of bridging modules, the HCI∧2 Framework is interoperable with other software frameworks including Psyclone and ActiveMQ. In addition to the core communication middleware, we also present the integrated development environment (IDE) of the HCI∧2 Framework. It provides a complete graphical environment to support every step in a typical MHCI system development process, including module development, debugging, packaging, and management, as well as the whole system management and testing. The quantitative evaluation indicates that our framework outperforms other similar tools in terms of average message latency and maximum data throughput under a typical single PC scenario. To demonstrate HCI∧2 Framework's capabilities in integrating heterogeneous modules, we present several example modules working with a variety of hardware and software. We also present an example of a full system developed using the proposed HCI∧2 Framework, which is called the CamGame system and represents a computer game based on hand-held marker(s) and low-cost camera(s).

Hidden conditional random fields (HCRFs) are discriminative latent variable models that have been shown to successfully learn the hidden structure of a given classification problem (provided an appropriate validation of the number of hidden states). In this brief, we present the infinite HCRF (iHCRF), which is a nonparametric model based on hierarchical Dirichlet processes and is capable of automatically learning the optimal number of hidden states for a classification task. We show how we learn the model hyperparameters with an effective Markov-chain Monte Carlo sampling technique, and we explain the process that underlines our iHCRF model with the Restaurant Franchise Rating Agencies analogy. We show that the iHCRF is able to converge to a correct number of represented hidden states, and outperforms the best finite HCRFs--chosen via cross-validation--for the difficult tasks of recognizing instances of agreement, disagreement, and pain. Moreover, the iHCRF manages to achieve this performance in significantly less total training, validation, and testing time.

In this paper we propose new binary pattern features for use in the problem of 3D facial action unit (AU) detection. Two representations of 3D facial geometries are employed, the depth map and the Azimuthal Projection Distance Image (APDI). To these the traditional Local Binary Pattern is applied, along with Local Phase Quantisation, Gabor filters and Monogenic filters, followed by the binary pattern feature extraction method. Feature vectors are formed for each feature type through concatenation of histograms formed from the resulting binary numbers. Feature selection is then performed using a two-stage GentleBoost approach. Finally, we apply Support Vector Machines as classifiers for detection of each AU. This system is tested in two ways. First we perform 10-fold cross-validation on the Bosphorus database, and then we perform cross-database testing by training on this database and then testing on apex frames from the D3DFACS database, achieving promising results in both.

Automated facial expression recognition has received increased attention over the past two decades. Existing works in the field usually do not encode either the temporal evolution or the intensity of the observed facial displays. They also fail to jointly model multidimensional (multi-class) continuous facial behaviour data; binary classifiers - one for each target basic-emotion class - are used instead. In this paper, intrinsic topology of multidimensional continuous facial affect data is first modeled by an ordinal manifold. This topology is then incorporated into the Hidden Conditional Ordinal Random Field (H-CORF) framework for dynamic ordinal regression by constraining H-CORF parameters to lie on the ordinal manifold. The resulting model attains simultaneous dynamic recognition and intensity estimation of facial expressions of multiple emotions. To the best of our knowledge, the proposed method is the first one to achieve this on both deliberate as well as spontaneous facial affect data.

In this study, we investigate an audiovisual approach for classification of vocal outbursts (non-linguistic vocalisations) in noisy conditions using Long Short-Term Memory (LSTM) Recurrent Neural Networks and Support Vector Machines. Fusion of geometric shape features and acoustic low-level descriptors is performed on the feature level. Three different types of acoustic noise are considered: babble, office and street noise. Experiments are conducted on every noise type to asses the benefit of the fusion in each case. As database for evaluations serves the INTERSPEECH 2010 Paralinguistic Challenge's Audiovisual Interest Corpus of human-to-human natural conversation. The results show that even when training is performed on noise corrupted audio which matches the test conditions the addition of visual features is still beneficial.

This paper proposes a new feature descriptor, local normal binary patterns (LNBPs), which is exploited for detection of facial action units (AUs). After LNBPs have been employed to form descriptor vectors, which capture the detailed shape of the action, feature selection is performed via a Gentle-Boost (GB) algorithm, and support vector machines (SVMs) are trained to detect each AU. This process was tested on the Bosphorus database, alongside the same test using 3D local binary pattern (3DLBP) descriptors which apply the LBP operator to the depth map of the face. LNBP descriptors were demonstrated to outperform 3DLBPs in detection of many individual AUs. Finally, feature fusion was used to combine the benefits of the 3DLBPs and each of the LNBP descriptors, with the best result achieving a mean ROC AuC of 96.35.

The second international Audio/Visual Emotion Challenge and Workshop 2012 (AVEC 2012) is introduced shortly. 34 teams from 12 countries signed up for the Challenge. The SEMAINE database serves for prediction of four-dimensional continuous affect in audio and video. For the eligible participants, final scores for the Fully-Continuous Sub-Challenge ranged between a correlation coefficient between gold standard and prediction of 0.174 and 0.456, and for Word-Level Sub-Challenge between 0.113 and 0.280.

We present the second Audio-Visual Emotion recognition Challenge and workshop (AVEC 2012), which aims to bring together researchers from the audio and video analysis communities around the topic of emotion recognition. The goal of the challenge is to recognise four continuously valued affective dimensions: arousal, expectancy, power, and valence. There are two sub-challenges: in the Fully Continuous Sub-Challenge participants have to predict the values of the four dimensions at every moment during the recordings, while for the Word-Level Sub-Challenge a single prediction has to be given per word uttered by the user. This paper presents the challenge guidelines, the common data used, and the performance of the baseline system on the two tasks.

There is evidence in neuroscience indicating that prediction of spatial and temporal patterns in the brain plays a key role in perception. This has given rise to prediction-based fusion as a method of combining information from audio and visual modalities. Models are trained on a per-class basis, to learn the mapping from one feature-space to another. When presented with unseen data, each model predicts the respective feature-sets using its learnt mapping, and the prediction error is combined within each class. The model which best describes the audiovisual relationship (by having the lowest combined prediction error) provides its label to the input data. Previous studies have only used neural networks to evaluate this method of combining modalities - this paper extends this to other learning methods, including Long Short-Term Memory recurrent neural networks (LSTMs), Support Vector Machines (SVMs), Relevance Vector Machines (RVMs), and Gaussian Processes (GPs). Our results on cross-database experiments on nonlinguistic vocalisation recognition show that feature-prediction significantly outperforms feature-fusion for neural networks, LSTMs, and GPs, while performance on SVMs and RVMs is more ambiguous and neither model gains an absolute advantage over the other.

Tags are an effective form of metadata which help users to locate and browse multimedia content of interest. Tags can be generated by users (user-generated explicit tags), automatically from the content (content-based tags), or assigned automatically based on non-verbal behavioral reactions of users to multimedia content (implicit human-centered tags). This paper discusses the definition and applications of implicit human-centered tagging. Implicit tagging is an effortless process by which content is tagged based on users' spontaneous reactions. It is a novel but growing research topic which is attracting more attention with the growing availability of built-in sensors. This paper discusses the state of the art in this novel field of research and provides an overview of publicly available relevant databases and annotation tools. We finally discuss in detail challenges and opportunities in the field.

We introduce the notion of subspace learning from image gradient orientations for appearance-based object recognition. As image data are typically noisy and noise is substantially different from Gaussian, traditional subspace learning from pixel intensities very often fails to estimate reliably the low-dimensional subspace of a given data population. We show that replacing pixel intensities with gradient orientations and the ℓ₂ norm with a cosine-based distance measure offers, to some extend, a remedy to this problem. Within this framework, which we coin Image Gradient Orientations (IGO) subspace learning, we first formulate and study the properties of Principal Component Analysis of image gradient orientations (IGO-PCA). We then show its connection to previously proposed robust PCA techniques both theoretically and experimentally. Finally, we derive a number of other popular subspace learning techniques, namely, Linear Discriminant Analysis (LDA), Locally Linear Embedding (LLE), and Laplacian Eigenmaps (LE). Experimental results show that our algorithms significantly outperform popular methods such as Gabor features and Local Binary Patterns and achieve state-of-the-art performance for difficult problems such as illumination and occlusion-robust face recognition. In addition to this, the proposed IGO-methods require the eigendecomposition of simple covariance matrices and are as computationally efficient as their corresponding ℓ₂ norm intensity-based counterparts. Matlab code for the methods presented in this paper can be found at http://ibug.doc.ic.ac.uk/resources.

SEMAINE has created a large audiovisual database as a part of an iterative approach to building Sensitive Artificial Listener (SAL) agents that can engage a person in a sustained, emotionally colored conversation. Data used to build the agents came from interactions between users and an "operator” simulating a SAL agent, in different configurations: Solid SAL (designed so that operators displayed an appropriate nonverbal behavior) and Semi-automatic SAL (designed so that users' experience approximated interacting with a machine). We then recorded user interactions with the developed system, Automatic SAL, comparing the most communicatively competent version to versions with reduced nonverbal skills. High quality recording was provided by five high-resolution, high-framerate cameras, and four microphones, recorded synchronously. Recordings total 150 participants, for a total of 959 conversations with individual SAL characters, lasting approximately 5 minutes each. Solid SAL recordings are transcribed and extensively annotated: 6-8 raters per clip traced five affective dimensions and 27 associated categories. Other scenarios are labeled on the same pattern, but less fully. Additional information includes FACS annotation on selected extracts, identification of laughs, nods, and shakes, and measures of user engagement with the automatic system. The material is available through a web-accessible database.

This paper describes a substantial effort to build a real-time interactive multimodal dialogue system with a focus on emotional and nonverbal interaction capabilities. The work is motivated by the aim to provide technology with competences in perceiving and producing the emotional and nonverbal behaviors required to sustain a conversational dialogue. We present the Sensitive Artificial Listener (SAL) scenario as a setting which seems particularly suited for the study of emotional and nonverbal behavior since it requires only very limited verbal understanding on the part of the machine. This scenario allows us to concentrate on nonverbal capabilities without having to address at the same time the challenges of spoken language understanding, task modeling, etc. We first report on three prototype versions of the SAL scenario in which the behavior of the Sensitive Artificial Listener characters was determined by a human operator. These prototypes served the purpose of verifying the effectiveness of the SAL scenario and allowed us to collect data required for building system components for analyzing and synthesizing the respective behaviors. We then describe the fully autonomous integrated real-time system we created, which combines incremental analysis of user behavior, dialogue management, and synthesis of speaker and listener behavior of a SAL character displayed as a virtual agent. We discuss principles that should underlie the evaluation of SAL-type systems. Since the system is designed for modularity and reuse and since it is publicly available, the SAL system has potential as a joint research tool in the affective computing research community.

MAHNOB-HCI is a multimodal database recorded in response to affective stimuli with the goal of emotion recognition and implicit tagging research. A multimodal setup was arranged for synchronized recording of face videos, audio signals, eye gaze data, and peripheral/central nervous system physiological signals. Twenty-seven participants from both genders and different cultural backgrounds participated in two experiments. In the first experiment, they watched 20 emotional videos and self-reported their felt emotions using arousal, valence, dominance, and predictability as well as emotional keywords. In the second experiment, short videos and images were shown once without any tag and then with correct or incorrect tags. Agreement or disagreement with the displayed tags was assessed by the participants. The recorded videos and bodily responses were segmented and stored in a database. The database is made available to the academic community via a web-based system. The collected data were analyzed and single modality and modality fusion results for both emotion recognition and implicit tagging experiments are reported. These results show the potential uses of the recorded modalities and the significance of the emotion elicitation protocol.

This paper presents a user-independent emotion recognition method with the goal of recovering affective tags for videos using electroencephalogram (EEG), pupillary response and gaze distance. We first selected 20 video clips with extrinsic emotional content from movies and online resources. Then, EEG responses and eye gaze data were recorded from 24 participants while watching emotional video clips. Ground truth was defined based on the median arousal and valence scores given to clips in a preliminary study using an online questionnaire. Based on the participants' responses, three classes for each dimension were defined. The arousal classes were calm, medium aroused, and activated and the valence classes were unpleasant, neutral, and pleasant. One of the three affective labels of either valence or arousal was determined by classification of bodily responses. A one-participant-out cross validation was employed to investigate the classification performance in a user-independent approach. The best classification accuracies of 68.5 percent for three labels of valence and 76.4 percent for three labels of arousal were obtained using a modality fusion strategy and a support vector machine. The results over a population of 24 participants demonstrate that user-independent emotion recognition can outperform individual self-reports for arousal assessments and do not underperform for valence assessments.

Social Signal Processing is the research domain aimed at bridging the social intelligence gap between humans and machines. This paper is the first survey of the domain that jointly considers its three major aspects, namely, modeling, analysis, and synthesis of social behavior. Modeling investigates laws and principles underlying social interaction, analysis explores approaches for automatic understanding of social exchanges recorded with different sensors, and synthesis studies techniques for the generation of social behavior via various forms of embodiment. For each of the above aspects, the paper includes an extensive survey of the literature, points to the most important publicly available resources, and outlines the most fundamental challenges ahead.

We propose an exact framework for online learning with a family of indefinite (not positive) kernels. As we study the case of nonpositive kernels, we first show how to extend kernel principal component analysis (KPCA) from a reproducing kernel Hilbert space to Krein space. We then formulate an incremental KPCA in Krein space that does not require the calculation of preimages and therefore is both efficient and exact. Our approach has been motivated by the application of visual tracking for which we wish to employ a robust gradient-based kernel. We use the proposed nonlinear appearance model learned online via KPCA in Krein space for visual tracking in many popular and difficult tracking scenarios. We also show applications of our kernel framework for the problem of face recognition.

Automatic facial expression recognition has been an active topic in computer science for over two decades, in particular facial action coding system action unit (AU) detection and classification of a number of discrete emotion states from facial expressive imagery. Standardization and comparability have received some attention; for instance, there exist a number of commonly used facial expression databases. However, lack of a commonly accepted evaluation protocol and, typically, lack of sufficient details needed to reproduce the reported individual results make it difficult to compare systems. This, in turn, hinders the progress of the field. A periodical challenge in facial expression recognition would allow such a comparison on a level playing field. It would provide an insight on how far the field has come and would allow researchers to identify new goals, challenges, and targets. This paper presents a meta-analysis of the first such challenge in automatic recognition of facial expressions, held during the IEEE conference on Face and Gesture Recognition 2011. It details the challenge data, evaluation protocol, and the results attained in two subchallenges: AU detection and classification of facial expression imagery in terms of a number of discrete emotion categories. We also summarize the lessons learned and reflect on the future of the field of facial expression recognition in general and on possible future challenges in particular.

Past work on automatic analysis of facial expressions has focused mostly on detecting prototypic expressions of basic emotions like happiness and anger. The method proposed here enables the detection of a much larger range of facial behavior by recognizing facial muscle actions [action units (AUs)] that compound expressions. AUs are agnostic, leaving the inference about conveyed intent to higher order decision making (e.g., emotion recognition). The proposed fully automatic method not only allows the recognition of 22 AUs but also explicitly models their temporal characteristics (i.e., sequences of temporal segments: neutral, onset, apex, and offset). To do so, it uses a facial point detector based on Gabor-feature-based boosted classifiers to automatically localize 20 facial fiducial points. These points are tracked through a sequence of images using a method called particle filtering with factorized likelihoods. To encode AUs and their temporal activation models based on the tracking data, it applies a combination of GentleBoost, support vector machines, and hidden Markov models. We attain an average AU recognition rate of 95.3% when tested on a benchmark set of deliberately displayed facial expressions and 72% when tested on spontaneous expressions.

This paper attempts to recognize spontaneous agreement and disagreement based only on nonverbal multi-modal cues. Related work has mainly used verbal and prosodic cues. We demonstrate that it is possible to correctly recognize agreement and disagreement without the use of verbal context (i.e. words, syntax). We propose to explicitly model the complex hidden dynamics of the multimodal cues using a sequential discriminative model, the Hidden Conditional Random Field (HCRF). In this paper, we show that the HCRF model is able to capture what makes each of these social attitudes unique. We present an efficient technique to analyze the concepts learned by the HCRF model and show that these coincide with the findings from social psychology regarding which cues are most prevalent in agreement and disagreement. Our experiments are performed on a spontaneous dataset of real televised debates. The HCRF model outperforms conventional approaches such as Hidden Markov Models and Support Vector Machines.

The automatic assessment of affect is mostly based on feature-level approaches, such as distances between facial points or prosodic and spectral information when it comes to audiovisual analysis. However, it is known and intuitive that behavioural events such as smiles, head shakes or laughter and sighs also bear highly relevant information regarding a subject's affective display. Accordingly, we propose a novel string-based prediction approach to fuse such events and to predict human affect in a continuous dimensional space. Extensive analysis and evaluation has been conducted using the newly released SEMAINE database of human-to-agent communication. For a thorough understanding of the obtained results, we provide additional benchmarks by more conventional feature-level modelling, and compare these and the string-based approach to fusion of signal-based features and string-based events. Our experimental results show that the proposed string-based approach is the best performing approach for automatic prediction of Valence and Expectation dimensions, and improves prediction performance for the other dimensions when combined with at least acoustic signal-based features.

Despite major advances within the affective computing research field, modelling, analysing, interpreting and responding to naturalistic human affective behaviour still remains as a challenge for automated systems as emotions are complex constructs with fuzzy boundaries and with substantial individual variations in expression and experience. Thus, a small number of discrete categories (e.g., happiness and sadness) may not reflect the subtlety and complexity of the affective states conveyed by such rich sources of information. Therefore, affective and behavioural computing researchers have recently invested increased effort in exploring how to best model, analyse, interpret and respond to the subtlety, complexity and continuity (represented along a continuum e.g., from -1 to +1) of affective behaviour in terms of latent dimensions (e.g., arousal, power and valence) and appraisals. Accordingly, this paper aims to present the current state of the art and the new challenges in automatic, dimensional and continuous analysis and synthesis of human emotional behaviour in an interdisciplinary perspective.

Recently developed appearance descriptors offer the opportunity for efficient and robust facial expression recognition. In this paper we investigate the merits of the family of local binary pattern descriptors for FACS Action-Unit (AU) detection. We compare Local Binary Patterns (LBP) and Local Phase Quantisation (LPQ) for static AU analysis. To encode facial expression dynamics, we extend the purely spatial representation LPQ to a dynamic texture descriptor which we call Local Phase Quantisation from Three Orthogonal Planes (LPQ-TOP), and compare this with the Local Binary Patterns from Three Orthogonal Planes (LBP-TOP). The efficiency of these descriptors is evaluated by a fully automatic AU detection system and tested on posed and spontaneous expression data collected from the MMI and SEMAINE databases. Results show that the systems based on LPQ achieve higher accuracy rate than those using LBP, and that the systems that utilise dynamic appearance descriptors outperform those that use static appearance descriptors. Overall, our proposed LPQ-TOP method outperformed all other tested methods.

We introduce a fast and robust subspace-based approach to appearance-based object tracking. The core of our approach is based on Fast Robust Correlation (FRC), a recently proposed technique for the robust estimation of large translational displacements. We show how the basic principles of FRC can be naturally extended to formulate a robust version of Principal Component Analysis (PCA) which can be efficiently implemented incrementally and therefore is particularly suitable for robust real-time appearance-based object tracking. Our experimental results demonstrate that the proposed approach outperforms other state-of-the-art holistic appearance-based trackers on several popular video sequences.

Many problems in machine learning and computer vision consist of predicting multi-dimensional output vectors given a specific set of input features. In many of these problems, there exist inherent temporal and spacial dependencies between the output vectors, as well as repeating output patterns and input-output associations, that can provide more robust and accurate predictors when modelled properly. With this intrinsic motivation, we propose a novel Output-Associative Relevance Vector Machine (OA-RVM) regression framework that augments the traditional RVM regression by being able to learn non-linear input and output dependencies. Instead of depending solely on the input patterns, OA-RVM models output structure and covariances within a predefined temporal window, thus capturing past, current and future context. As a result, output patterns manifested in the training data are captured within a formal probabilistic framework, and subsequently used during inference. As a proof of concept, we target the highly challenging problem of dimensional and continuous prediction of emotions from naturalistic facial expressions. We demonstrate the advantages of the proposed OA-RVM regression by performing both subject-dependent and subject-independent experiments using the SAL database. The experimental results show that OA-RVM regression outperforms the traditional RVM and SVM regression approaches in prediction accuracy, generating more robust and accurate models.

Recent evidence in neuroscience support the theory that prediction of spatial and temporal patterns in the brain plays a key role in human actions and perception. Inspired by these findings, a system that discriminates laughter from speech by modeling the spatial and temporal relationship between audio and visual features is presented. The underlying assumption is that this relationship is different between speech and laughter. Neural networks are trained which learn the audio-to-visual and visual-to-audio feature mapping together with the time evolution of audio and visual features for both classes. Classification of a new frame / sequence is performed via prediction. All the networks produce a prediction of the expected audio / visual features and their prediction errors are combined for each class. The model which best describes the audiovisual feature relationship, i.e., results in the lowest prediction error, provides its label to the input frame / sequence. Using 4 different datasets, the proposed system is compared to standard feature-level fusion on cross-database experiments. In almost all test cases, prediction-based classification outperforms feature-level fusion. Similar conclusion are drawn when adding artificial feature-level noise to the datasets.

In this paper we propose a method that exploits 3D motion-based features between frames of 3D facial geometry sequences for dynamic facial expression recognition. An expressive sequence is modeled to contain an onset followed by an apex and an offset. Feature selection methods are applied in order to extract features for each of the onset and offset segments of the expression. These features are then used to train a Hidden Markov Model in order to model the full temporal dynamics of the expression. The proposed fully automatic system was tested in a subset of the BU-4DFE database for the recognition of happiness, anger and surprise. Comparisons with a similar system based on the motion extracted from facial intensity images was also performed. The attained results suggest that the use of the 3D information does indeed improve the recognition accuracy when compared to the 2D data.

This demonstration aims to showcase the recently completed SEMAINE system. The SEMAINE system is a publicly available, fully autonomous Sensitive Artificial Listeners (SAL) system that consists of virtual dialog partners based on audiovisual analysis and synthesis (see http://semaine.opendfki.de/wiki). The system runs in real-time, and combines incremental analysis of user behavior, dialog management, and synthesis of speaker and listener behavior of a SAL character, displayed as a virtual agent. The SAL characters intend to engage the user in a conversation by paying attention to the user's emotions and nonverbal expressions. The characters have their own emotionally defined personality. During an interaction, the characters attempt to create an emotional workout for the user by drawing her/him towards their dominant emotion, through a combination of verbal and nonverbal expressions.

In this paper, we present a novel software tool designed and implemented to simplify the development process of Multimodal Human-Computer Interaction (MHCI) systems. This tool, which is called the HCI^2 Workbench, exploits a Publish / Subscribe (P/S) architecture to facilitate efficient and reliable inter-module data communication and runtime system management. In addition, through a combination of SDK, software tools, and standardized description / configuration file semantics, the HCI^2 Workbench provides an easy-to-follow procedure for developing highly flexible and reusable modules. Moreover, the HCI^2 Workbench features system persistence and portability by using standardized module packaging method and system configuration files. Last but not least, usability was another major concern. Unlike other similar tool, including Psyclone and ActiveMQ, the HCI^2 Workbench provides a complete graphical environment to support every step in a typical MHCI system development process, including module program development and debugging, module packaging, module management, system configuration, module and system testing, in a convenient and intuitive manner. To help demonstrating the HCI^2 Workbench, we also present a readily-applicable system developed using our tool. This open-source demo system, which is called the CamGame, consists of an interactive system allowing users to play a computer game using hand-held marker(s) and low-cost camera(s) instead of keyboard and mouse.

We introduce the notion of Principal Component Analysis (PCA) of image gradient orientations. As image data is typically noisy, but noise is substantially different from Gaussian, traditional PCA of pixel intensities very often fails to estimate reliably the low-dimensional subspace of a given data population. We show that replacing intensities with gradient orientations and the ℓ2norm with a cosine-based distance measure offers, to some extend, a remedy to this problem. Our scheme requires the eigen-decomposition of a covariance matrix and is as computationally efficient as standard ℓ2intensity-based PCA. We demonstrate some of its favorable properties for the application of face recognition.

We investigate classification of non-linguistic vocalisations with a novel audiovisual approach and Long Short-Term Memory (LSTM) Recurrent Neural Networks as highly successful dynamic sequence classifiers. As database of evaluation serves this year's Paralinguistic Challenge's Audiovisual Interest Corpus of human-to-human natural conversation. For video-based analysis we compare shape and appearance based features. These are fused in an early manner with typical audio descriptors. The results show significant improvements of LSTM networks over a static approach based on Support Vector Machines. More important, we can show a significant gain in performance when fusing audio and visual shape features.

Given the facial points extracted from an image of a face in an arbitrary pose, the goal of facial-point-based head-pose normalization is to obtain the corresponding facial points in a predefined pose (e.g., frontal). This involves inference of complex and high-dimensional mappings due to the large number of the facial points employed, and due to differences in head-pose and facial expression. Most regression-based approaches for learning such mappings focus on modeling correlations only between the inputs (i.e., the facial points in a non-frontal pose) and the outputs (i.e., the facial points in the frontal pose), but not within the inputs and the outputs of the model. This makes these models prone to errors due to noise and outliers in test data, often resulting in anatomically impossible facial configurations formed by their predictions. To address this, we propose Shape-constrained Gaussian Process (SC-GP) regression for facial-point-based head-pose normalization. Specifically, a deformable face-shape model is used to learn a face-shape prior, which is placed on both the input and the output of GP regression in order to constrain the model predictions to anatomically feasible facial configurations. Our extensive experiments on both synthetic and real image data show that the proposed approach generalizes well across poses and handles successfully noise and outliers in test data. In addition, the proposed model outperforms previously proposed approaches to facial-point-based head-pose normalization.

We propose a correlation-based approach to parametric object alignment particularly suitable for face analysis applications which require efficiency and robustness against occlusions and illumination changes. Our algorithm registers two images by iteratively maximizing their correlation coefficient using gradient ascent. We compute this correlation coefficient from complex gradients which capture the orientation of image structures rather than pixel intensities. The maximization of this gradient correlation coefficient results in an algorithm which is as computationally efficient as ℓ2norm-based algorithms, can be extended within the inverse compositional framework (without the need for Hessian recomputation) and is robust to outliers. To the best of our knowledge, no other algorithm has been proposed so far having all three features. We show the robustness of our algorithm for the problem of face alignment in the presence of occlusions and non-uniform illumination changes. The code that reproduces the results of our paper can be found at http://ibug.doc.ic.ac.uk/resources.

This paper investigates dimensional emotion prediction and classification from naturalistic facial expressions. Similarly to many pattern recognition problems, dimensional emotion classification requires generating multi-dimensional outputs. To date, classification for valence and arousal dimensions has been done separately, assuming that they are independent. However, various psychological findings suggest that these dimensions are correlated. We therefore propose a novel, multi-layer hybrid framework for emotion classification that is able to model inter-dimensional correlations. Firstly, we derive a novel geometric feature set based on the (a)symmetric spatio-temporal characteristics of facial expressions. Subsequently, we use the proposed feature set to train a multi-layer hybrid framework composed of a tem- poral regression layer for predicting emotion dimensions, a graphical model layer for modeling valence-arousal correlations, and a final classification and fusion layer exploiting informative statistics extracted from the lower layers. This framework (i) introduces the Auto-Regressive Coupled HMM (ACHMM), a graphical model specifically tailored to accommodate not only inter-dimensional correlations but also to exploit the internal dynamics of the actual observations, and (ii) replaces the commonly used Maximum Likelihood principle with a more robust final classification and fusion layer. Subject-independent experimental validation, performed on a naturalistic set of facial expressions, demonstrates the effectiveness of the derived feature set, and the robustness and flexibility of the proposed framework.

The ability to understand social signals of a person we are communicating with is the core of social intelligence. Social Intelligence is a facet of human intelligence that has been argued to be indispensable and perhaps the most important for success in life. At the same time, human-centric multimedia applications for humans and about humans are becoming increasingly important. 3D modeled human-objects, like bodies, heads and faces are exploited for animation, security, and human computer interaction, while three dimensional motion of arms, legs and local body features is used for more complete human gesture, activity and behavior analysis. The Joint Human Gesture and Behavior Understanding (J-HGBU) workshop event consists of two parts focusing on these complementary challenges: the Workshop on Multimedia Access to 3D Human Objects (MA3HO'11) and the Workshop on Social Signal Processing (SSPW'11).

Social signal processing has the ambitious goal of bridging the social intelligence gap between computers and humans. Nowadays, computers are not only the new interaction partners of humans, but also a privileged interaction medium for social exchange between humans. Consequently, enhancing machine abilities to interpret and reproduce social signals is a crucial requirement for improving computer-mediated communication and interaction. Furthermore, automated analysis of such signals creates a host of new applications and improvements to existing applications. The study of social signals benefits a wide range of domains, including human-computer interaction, interaction design, entertainment technology, ambient intelligence, health-care, and psychology. This paper briefly introduces the field and surveys its latest developments.

During face-to-face interpersonal interaction, people have a tendency to mimic each other. People not only mimic postures, mannerisms, moods or emotions, but they also mimic several speech-related behaviors. In this paper we describe how visual and vocal behavioral information expressed between two interlocutors can be used to detect and identify visual and vocal mimicry. We investigate expressions of mimicry and aim to learn more about in which situation and to what extent mimicry occurs. The observable effects of mimicry can be explored by representing and recognizing mimicry using visual and vocal features. In order to automatically analyze how to extract and integrate this behavioral information into a multimodal mimicry detection framework for improving affective computing, this paper addresses the main challenge: mimicry representation in terms of optimal behavioral feature extraction and automatic integration in both audio and video modalities.

Past research in analysis of human affect has focused on recognition of prototypic expressions of six basic emotions based on posed data acquired in laboratory settings. Recently, there has been a shift toward subtle, continuous, and context-specific interpretations of affective displays recorded in naturalistic and real-world settings, and toward multimodal analysis and recognition of human affect. Converging with this shift, this paper presents, to the best of our knowledge, the first approach in the literature that: 1) fuses facial expression, shoulder gesture, and audio cues for dimensional and continuous prediction of emotions in valence and arousal space, 2) compares the performance of two state-of-the-art machine learning techniques applied to the target problem, the bidirectional Long Short-Term Memory neural networks (BLSTM-NNs), and Support Vector Machines for Regression (SVR), and 3) proposes an output-associative fusion framework that incorporates correlations and covariances between the emotion dimensions. Evaluation of the proposed approach has been done using the spontaneous SAL data from four subjects and subject-dependent leave-one-sequence-out cross validation. The experimental results obtained show that: 1) on average, BLSTM-NNs outperform SVR due to their ability to learn past and future context, 2) the proposed output-associative fusion framework outperforms feature-level and model-level fusion by modeling and learning correlations and patterns between the valence and arousal dimensions, and 3) the proposed system is well able to reproduce the valence and arousal ground truth obtained from human coders.

In this paper we address the problem of localization and recognition of human activities in unsegmented image sequences. The main contribution of the proposed method is the use of an implicit representation of the spatiotemporal shape of the activity which relies on the spatiotemporal localization of characteristic ensembles of feature descriptors. Evidence for the spatiotemporal localization of the activity is accumulated in a probabilistic spatiotemporal voting scheme. The local nature of the proposed voting framework allows us to deal with multiple activities taking place in the same scene, as well as with activities in the presence of clutter and occlusion. We use boosting in order to select characteristic ensembles per class. This leads to a set of class specific codebooks where each codeword is an ensemble of features. During training, we store the spatial positions of the codeword ensembles with respect to a set of reference points, as well as their temporal positions with respect to the start and end of the action instance. During testing, each activated codeword ensemble casts votes concerning the spatiotemporal position and extend of the action, using the information that was stored during training. Mean Shift mode estimation in the voting space provides the most probable hypotheses concerning the localization of the subjects at each frame, as well as the extend of the activities depicted in the image sequences. We present classification and localization results for a number of publicly available datasets, and for a number of sequences where there is a significant amount of clutter and occlusion.

Past research on automatic laughter classification/detection has focused mainly on audio-based approaches. Here we present an audiovisual approach to distinguishing laughter from speech, and we show that integrating the information from audio and video channels may lead to improved performance over single-modal approaches. Both audio and visual channels consist of two streams (cues), facial expressions and head pose for video and cepstral and prosodic features for audio. Two types of experiments were performed: 1) subject-independent cross-validation on the AMI dataset and 2) cross-database experiments on the AMI and SAL datasets. We experimented with different combinations of cues with the most informative being the combination of facial expressions, cepstral, and prosodic features. Our results suggest that the performance of the audiovisual approach is better on average than single-modal approaches. The addition of visual information produces better results when it comes to female subjects. When the training conditions are less diverse in terms of head movements than the testing conditions (training on the SAL dataset, testing on the AMI dataset), then no improvement was observed with the addition of visual information. On the other hand, when the training conditions are similar (cross validation on the AMI dataset), or more diverse (training on the AMI dataset, testing on the SAL dataset), in terms of head movements than is the case in the testing conditions, an absolute increase of about 3% in the F1 rate for laughter is reported when visual information is added to audio information.

Finding fiducial facial points in any frame of a video showing rich naturalistic facial behaviour is an unsolved problem. Yet this is a crucial step for geometric-feature-based facial expression analysis, and methods that use appearance-based features extracted at fiducial facial point locations. In this paper we present a method based on a combination of Support Vector Regression and Markov Random Fields to drastically reduce the time needed to search for a point's location and increase the accuracy and robustness of the algorithm. Using Markov Random Fields allows us to constrain the search space by exploiting the constellations that facial points can form. The regressors on the other hand learn a mapping between the appearance of the area surrounding a point and the positions of these points, which makes detection of the points very fast and can make the algorithm robust to variations of appearance due to facial expression and moderate changes in head pose. The proposed point detection algorithm was tested on 1855 images, the results of which showed we outperform current state of the art point detectors.

In this study, a system that discriminates laughter from speech by modelling the relationship between audio and visual features is presented. The underlying assumption is that this relationship is different between speech and laughter. Neural networks are trained which learn the audio-to-visual and visual-to-audio features mapping for both classes. Classification of a new frame is performed via prediction. All the networks produce a prediction of the expected audio/visual features and the network with the best prediction, i.e., the model which best describes the audiovisual feature relationship, provides its label to the input frame. When trained on a simple dataset and tested on a hard dataset, the proposed approach outperforms audiovisual feature-level fusion, resulting in a 10.9% and 6.4% absolute increase in the F1 rate for laughter and classification rate, respectively. This indicates that classification based on prediction can produce a good model even when the available dataset is not challenging enough.

We have recorded a new corpus of emotionally coloured conversations. Users were recorded while holding conversations with an operator who adopts in sequence four roles designed to evoke emotional reactions. The operator and the user are seated in separate rooms; they see each other through teleprompter screens, and hear each other through speakers. To allow high quality recording, they are recorded by five high-resolution, high framerate cameras, and by four microphones. All sensor information is recorded synchronously, with an accuracy of 25 μs. In total, we have recorded 20 participants, for a total of 100 character conversational and 50 non-conversational recordings of approximately 5 minutes each. All recorded conversations have been fully transcribed and annotated for five affective dimensions and partially annotated for 27 other dimensions. The corpus has been made available to the scientific community through a web-accessible database.

This paper focuses on audio-visual (using facial expression, shoulder and audio cues) classification of spontaneous affect, utilising generative models for classification (i) in terms of Maximum Likelihood Classification with the assumption that the generative model structure in the classifier is correct, and (ii) Likelihood Space Classification with the assumption that the generative model structure in the classifier may be incorrect, and therefore, the classification performance can be improved by projecting the results of generative classifiers onto likelihood space, and then using discriminative classifiers. Experiments are conducted by utilising Hidden Markov Models for single cue classification, and 2 and 3-chain coupled Hidden Markov Models for fusing multiple cues and modalities. For discriminative classification, we utilise Support Vector Machines. Results show that Likelihood Space Classification improves the performance (91.76%) of Maximum Likelihood Classification (79.1%). Thereafter, we introduce the concept of fusion in the likelihood space, which is shown to outperform the typically used model-level fusion, attaining a classification accuracy of 94.01% and further improving all previous results.

We present a regression-based scheme for multi-view facial expression recognition based on 2D geometric features. We address the problem by mapping facial points (e.g. mouth corners) from non-frontal to frontal view where further recognition of the expressions can be performed using a state-of-the-art facial expression recognition method. To learn the mapping functions we investigate four regression models: Linear Regression (LR), Support Vector Regression (SVR), Relevance Vector Regression (RVR) and Gaussian Process Regression (GPR). Our extensive experiments on the CMU Multi-PIE facial expression database show that the proposed scheme outperforms view-specific classifiers by utilizing considerably less training data.

The classification of sequences requires the combination of information from different time points. In this paper the detection of facial expressions is considered. Experiments on the detection of certain facial muscle activations in videos show that it is not always required to model the sequences fully, but that the presence of specific frames (the concept frame) can be sufficient for a reliable detection of certain facial expression classes. For the detection of these concept frames a standard classifier is often sufficient, although a more advanced clustering approach performs better in some cases.

The Workshop on Social Signal Processing (SSPW) is the yearly event of the Social Signal Processing Network (EU-FP7 SSPNet project). This year's workshop programme consists of 4 premium Key Note Talks by Jeff Cohn, Alex Pentland. Justine Cassell, and Toyoaki Nishida, an oral session with 4 presentations, a poster session with 7 posters, and a panel session where the panelists will be the Key Note Speakers and the workshop organizers.

In this work, we propose a dynamic texture-based approach to the recognition of facial Action Units (AUs, atomic facial gestures) and their temporal models (i.e., sequences of temporal segments: neutral, onset, apex, and offset) in near-frontal-view face videos. Two approaches to modeling the dynamics and the appearance in the face region of an input video are compared: an extended version of Motion History Images and a novel method based on Nonrigid Registration using Free-Form Deformations (FFDs). The extracted motion representation is used to derive motion orientation histogram descriptors in both the spatial and temporal domain. Per AU, a combination of discriminative, frame-based GentleBoost ensemble learners and dynamic, generative Hidden Markov Models detects the presence of the AU in question and its temporal segments in an input image sequence. When tested for recognition of all 27 lower and upper face AUs, occurring alone or in combination in 264 sequences from the MMI facial expression database, the proposed method achieved an average event recognition accuracy of 89.2 percent for the MHI method and 94.3 percent for the FFD method. The generalization performance of the FFD method has been tested using the Cohn-Kanade database. Finally, we also explored the performance on spontaneous expressions in the Sensitive Artificial Listener data set.

Applications such as surveillance and human motion capture require high-bandwidth recording from multiple cameras. Furthermore, the recent increase in research on sensor fusion has raised the demand on synchronization accuracy between video, audio and other sensor modalities. Previously, capturing synchronized, high resolution video from multiple cameras required complex, inflexible and expensive solutions. Our experiments show that a single PC, built from contemporary low-cost computer hardware, could currently handle up to 470MB/s of input data. This allows capturing from 18 cameras of 780x580pixels at 60fps each, or 36 cameras at 30fps. Furthermore, we achieve accurate synchronization between audio, video and additional sensors, by recording audio together with sensor trigger- or timestamp signals, using a multi-channel audio input. In this way, each sensor modality can be captured with separate software and hardware, allowing maximal flexibility with minimal cost.

This paper presents the results of an empirical study suggesting that, while laughter is a very good indicator of amusement, the kind of laughter (unvoiced laughter vs.voiced laughter) is correlated with the mirth of laughter and could potential be used to judge the actual hilarity of the stimulus joke. For this study, an automated method for audiovisual analysis of laugher episodes exhibited while watching movie clips or observing the behaviour of a conversational agent has been developed. The audio and visual features, based on spectral properties of the acoustic signal and facial expressions respectively, have been integrated using feature level fusion, resulting in a multimodal approach to distinguishing voiced laughter from unvoiced laughter and speech. The classification accuracy of such a system tested on spontaneous laughter episodes is 74%. Finally, preliminary results are presented which provide evidence that unvoiced laughter can be interpreted as less gleeful than voiced laughter and consequently the detection of those two types of laughter can be used to label multimedia content as little funny or very funny respectively.

This paper provides a general introduction to the concept of implicit human-centered tagging (IHCT) - the automatic extraction of tags from nonverbal behavioral feedback of media users. The main idea behind IHCT is that nonverbal behaviors displayed when interacting with multimedia data (e.g., facial expressions, head nods, etc.) provide information useful for improving the tag sets associated with the data. As such behaviors are displayed naturally and spontaneously, no effort is required from the users, and this is why the resulting tagging process is said to be "implicit". Tags obtained through IHCT are expected to be more robust than tags associated with the data explicitly, at least in terms of: generality (they make sense to everybody) and statistical reliability (all tags will be sufficiently represented). The paper discusses these issues in detail and provides an overview of pioneering efforts in the field.

Human nonverbal behavior recognition from multiple cues and modalities has attracted a lot of interest in recent years. Despite the interest, many research questions, including the type of feature representation, choice of static vs. dynamic classification schemes, the number and type of cues or modalities to use, and the optimal way of fusing these, remain open research questions. This paper compares frame-based vs window-based feature representation and employs static vs. dynamic classification schemes for two distinct problems in the field of automatic human nonverbal behavior analysis: multicue discrimination between posed and spontaneous smiles from facial expressions, head and shoulder movements, and audio-visual discrimination between laughter and speech. Single cue and single modality results are compared to multicue and multimodal results by employing Neural Networks, Hidden Markov Models (HMMs), and 2- and 3-chain coupled HMMs. Subject independent experimental evaluation shows that: 1) both for static and dynamic classification, fusing data coming from multiple cues and modalities proves useful to the overall task of recognition, 2) the type of feature representation appears to have a direct impact on the classification performance, and 3) static classification is comparable to dynamic classification both for multicue discrimination between posed and spontaneous smiles, and audio-visual discrimination between laughter and speech.

This paper describes a software framework we designed and implemented for the development and research in the area of multimodal human-computer interface. The proposed framework is based on publish/subscribe architecture, which allows developers and researchers to conveniently configure, test and expand their system in a modular and incremental manner. In order to achieve reliable and efficient data transport between modules while still providing a high degree of system flexibility, the framework uses a shared-memory based data transport protocol for message delivery together with a TCP based system management protocol to maintain the integrity of system structure at runtime. The framework is delivered as a communication middleware, providing a basic system manager and well-documented, easy-to-use and open source C++ SDKs supporting both module development and server extension. The experimental comparison between the proposed framework and other similar tools available to the community indicates that our framework greatly outperforms the others in terms of average message latency, maximum data throughput and CPU consumption level, especially in heavy workload scenarios. To demonstrate the performance of our framework in real world applications, we have built a demo system which is used to detect faces and facial feature points in real-time captured video. The result shows our framework is capable of delivering some tens of megabytes of data per second effectively and efficiently even under tight resource constraint.

Automated analysis of human affective behavior has attracted increasing attention from researchers in psychology, computer science, linguistics, neuroscience, and related disciplines. However, the existing methods typically handle only deliberately displayed and exaggerated expressions of prototypical emotions despite the fact that deliberate behaviour differs in visual appearance, audio profile, and timing from spontaneously occurring behaviour. To address this problem, efforts to develop algorithms that can process naturally occurring human affective behaviour have recently emerged. Moreover, an increasing number of efforts are reported toward multimodal fusion for human affect analysis including audiovisual fusion, linguistic and paralinguistic fusion, and multi-cue visual fusion based on facial expressions, head movements, and body gestures. This paper introduces and surveys these recent advances. We first discuss human emotion perception from a psychological perspective. Next we examine available approaches to solving the problem of machine understanding of human affective behavior, and discuss important issues like the collection and availability of training and test data. We finally outline some of the scientific and engineering challenges to advancing human affect sensing technology.

The seven articles in this special issue focus on human computing. Most focus on two challenging issues in human computing, namely, machine analysis of human behavior in group interactions and context-sensitive modeling.

In this paper we propose an appearance-based approach to recognition of facial action units (AUs) and their temporal segments in frontal-view face videos. Non-rigid registration using free-form deformations is used to determine motion in the face region of an input video. The extracted motion fields are then used to derive motion histogram descriptors. Per AU, a combination of ensemble learners and hidden Markov models detects the presence of the AU in question and its temporal segment in each frame of an input sequence. When tested for recognition of all 27 lower and upper face AUs, occurring alone or in combination in 264 sequences from the MMI facial expression database, an average sequence classification rate of 94.3% was achieved.

We propose to demonstrate the emotionally aware painting fool, a novel system that combines a machine vision system able to recognise emotions with a non-photorealistic rendering (NPR) system to automatically produce portraits of the sitter in an emotionally enhanced style. During the demonstration, the vision system records a short video clip of a person showing a basic emotion. The system then analyses this video clip, locating facial features and tracking their motion. Using this tracking data, the system analyses which emotion was expressed and what the temporal dynamics of the expression were. This information is then passed to the NPR software. The detected emotion is used to choose appropriate (simulated) art materials, colour palettes, abstraction methods and painting styles, so that the rendered image may heighten the emotion being expressed. The live demonstration shows how each element of the emotionally aware painting fool functions and produces a portrait in approximately 7 minutes.

Past research on automatic laughter detection has focused mainly on audio-based detection. Here we present an audiovisual approach to distinguishing laughter from speech and we show that integrating the information from audio and video leads to an improved reliability of audiovisual approach in comparison to single-modal approaches. We also investigated the level at which audiovisual information should be fused for the best performance. When tested on 96 audiovisual sequences depicting spontaneously displayed (as opposed to posed) laughter and speech episodes, the proposed audiovisual feature-level approach achieved a 86.9% recall rate with 76.7% precision.

Previous research on automatic laughter detection has mainly been focused on audio-based detection. In this study we present an audio-visual approach to distinguishing laughter from speech based on temporal features and we show that integrating the information from audio and video channels leads to improved performance over single-modal approaches. Static features are extracted on an audio/video frame basis and then combined with temporal features extracted over a temporal window, describing the evolution of static features over time. The use of several different temporal features has been investigated and it has been shown that the addition of temporal information results in an improved performance over utilizing static information only. It is common to use a fixed set of temporal features which implies that all static features will exhibit the same behaviour over a temporal window. However, this does not always hold and we show that when AdaBoost is used as a feature selector, different temporal features for each static feature are selected, i.e., the temporal evolution of each static feature is described by different statistical measures. When tested on 96 audiovisual sequences, depicting spontaneously displayed (as opposed to posed) laughter and speech episodes, in a person independent way the proposed audiovisual approach achieves an F1 rate of over 89%.

Social Signal Processing (SSP) aims at the analysis of social behaviour in both Human-Human and Human-Computer interactions. SSP revolves around automatic sensing and interpretation of social signals, complex aggregates of nonverbal behaviours through which individuals express their attitudes towards other human (and virtual) participants in the current social context. As such, SSP integrates both engineering (speech analysis, computer vision, etc.) and human sciences (social psychology, anthropology, etc.) as it requires multimodal and multidisciplinary approaches. As of today, SSP is still in its early infancy, but the domain is quickly developing, and a growing number of works is appearing in the literature. This paper provides an introduction to nonverbal behaviour involved in social signals and a survey of the main results obtained so far in SSP. It also outlines possibilities and challenges that SSP is expected to face in the next years if it is to reach its full maturity.

The ability to understand and manage social signals of a person we are communicating with is the core of social intelligence. Social intelligence is a facet of human intelligence that has been argued to be indispensable and perhaps the most important for success in life. This paper argues that next-generation computing needs to include the essence of social intelligence - the ability to recognize human social signals and social behaviours like politeness, and disagreement - in order to become more effective and more efficient. Although each one of us understands the importance of social signals in everyday life situations, and in spite of recent advances in machine analysis of relevant behavioural cues like blinks, smiles, crossed arms, laughter, and similar, design and development of automated systems for Social Signal Processing (SSP) are rather difficult. This paper surveys the past efforts in solving these problems by a computer, it summarizes the relevant findings in social psychology, and it proposes a set of recommendations for enabling the development of the next generation of socially-aware computing.

In this paper we propose a tracking scheme specifically tailored for tracking human body parts in cluttered scenes. We model the background and the human skin using Gaussian mixture models and we combine these estimates to localize the features to be tracked. We further use these estimates to determine the pixels which belong to the background and those which belong to the subject's skin and we incorporate this information in the observation model of the used tracking scheme. For handling self-occlusion (i.e., when one body part occludes another), we incorporate the information about the direction of the observed motion into the propagation model of the used tracking scheme. We demonstrate that the proposed method outperforms the conventional condensation and auxiliary particle filtering when the hands and the head are the tracked body features. For the purposes of human body gesture recognition, we use a variant of the longest common subsequence algorithm (LCSS) in order to acquire a distance measure between the acquired trajectories and we use this measure in order to define new kernels for a relevance vector machine (RVM) classification scheme. We present results on real image sequences from a small database depicting people performing 15 aerobic exercises.

Automatic distinction between posed and spontaneous expressions is an unsolved problem. Previously cognitive sciences' studies indicated that the automatic separation of posed from spontaneous expressions is possible using the face modality alone. However, little is known about the information contained in head and shoulder motion. In this work, we propose to (i) distinguish between posed and spontaneous smiles by fusing the head, face, and shoulder modalities, (ii) investigate which modalities carry important information and how the information of the modalities relate to each other, and (iii) to which extent the temporal dynamics of these signals attribute to solving the problem. We use a cylindrical head tracker to track the head movements and two particle filtering techniques to track the facial and shoulder movements. Classification is performed by kernel methods combined with ensemble learning techniques. We investigated two aspects of multimodal fusion: the level of abstraction (i.e., early, mid-level, and late fusion) and the fusion rule used (i.e., sum, product and weight criteria). Experimental results from 100 videos displaying posed smiles and 102 videos displaying spontaneous smiles are presented. Best results were obtained with late fusion of all modalities when 94.0% of the videos were classified correctly.

Automated analysis of human affective behavior has attracted increasing attention from researchers in psychology, computer science, linguistics, neuroscience, and related disciplines. Promising approaches have been reported, including automatic methods for facial and vocal affect recognition. However, the existing methods typically handle only deliberately displayed and exaggerated expressions of prototypical emotions-despite the fact that deliberate behavior differs in visual and audio expressions from spontaneously occurring behavior. Recently efforts to develop algorithms that can process naturally occurring human affective behavior have emerged. This paper surveys these efforts. We first discuss human emotion perception from a psychological perspective. Next, we examine the available approaches to solving the problem of machine understanding of human affective behavior occurring in real-world settings. We finally outline some scientific and engineering challenges for advancing human affect sensing technology.

Automatic facial expression analysis is an important aspect of human machine interaction as the face is an important communicative medium. We use our face to signal interest, disagreement, intentions or mood through subtle facial motions and expressions. Work on automatic facial expression analysis can roughly be divided into the recognition of prototypic facial expressions such as the six basic emotional states and the recognition of atomic facial muscle actions (action units, AUs). Detection of AUs rather than emotions makes facial expression detection independent of culture-dependent interpretation, reduces the dimensionality of the problem and reduces the amount of training data required. Classic psychological studies suggest that humans consciously map AUs onto the basic emotion categories using a finite number of rules. On the other hand, recent studies suggest that humans recognize emotions unconsciously with a process that is perhaps best modeled by artificial neural networks (ANNs). This paper investigates these two claims. A comparison is made between detection of emotions directly from features vs. a two-step approach where we first detect AUs and use the AUs as input to either a rulebase or an ANN to recognize emotions. The results suggest that the two-step approach is possible with a small loss of accuracy and that biologically inspired classification techniques outperform those that approach the classification problem from a logical perspective, suggesting that biologically inspired classifiers are more suitable for computer-based analysis of facial behavior than logic inspired methods

This paper describes an intelligent system that we developed to support affective multimodal human-computer interaction (AMMHCI) where the user’s actions and emotions are modeled and then used to adapt the HCI and support the user in his or her activity. The proposed system, which we named Gaze-X, is based on sensing and interpretation of the human part of the computer’s context, known as W5+ (who, where, what, when, why, how). It integrates a number of natural human communicative modalities including speech, eye gaze direction, face and facial expression, and a number of standard HCI modalities like keystrokes, mouse movements, and active software identification, which, in turn, are fed into processes that provide decision making and adapt the HCI to support the user in his or her activity according to his or her preferences. To attain a system that can be educated, that can improve its knowledge and decision making through experience, we use case-based reasoning as the inference engine of Gaze-X. The utilized case base is a dynamic, incrementally self-organizing event-content-addressable memory that allows fact retrieval and evaluation of encountered events based upon the user preferences and the generalizations formed from prior input. To support concepts of concurrency, modularity/scalability, persistency, and mobility, Gaze-X has been built as an agent-based system where different agents are responsible for different parts of the processing. A usability study conducted in an office scenario with a number of users indicates that Gaze-X is perceived as effective, easy to use, useful, and affectively qualitative.

A widely accepted prediction is that computing will move to the background, weaving itself into the fabric of our everyday living spaces and projecting the human user into the foreground. If this prediction is to come true, then next generation computing, which we will call human computing, should be about anticipatory user interfaces that should be human-centered, built for humans based on human models. They should transcend the traditional keyboard and mouse to include natural, human-like interactive functions including understanding and emulating certain human behaviors such as affective and social signaling. This article discusses a number of components of human behavior, how they might be integrated into computers, and how far we are from realizing the front end of human computing, that is, how far are we from enabling computers to understand human behavior.

Past research on automatic facial expression analysis has focused mostly on the recognition of prototypic expressions of discrete emotions rather than on the analysis of dynamic changes over time, although the importance of temporal dynamics of facial expressions for interpretation of the observed facial behavior has been acknowledged for over 20 years. For instance, it has been shown that the temporal dynamics of spontaneous and volitional smiles are fundamentally different from each other. In this work, we argue that the same holds for the temporal dynamics of brow actions and show that velocity, duration, and order of occurrence of brow actions are highly relevant parameters for distinguishing posed from spontaneous brow actions. The proposed system for discrimination between volitional and spontaneous brow actions is based on automatic detection of Action Units (AUs) and their temporal segments (onset, apex, offset) produced by movements of the eyebrows. For each temporal segment of an activated AU, we compute a number of mid-level feature parameters including the maximal intensity, duration, and order of occurrence. We use Gentle Boost to select the most important of these parameters. The selected parameters are used further to train Relevance Vector Machines to determine per temporal segment of an activated AU whether the action was displayed spontaneously or volitionally. Finally, a probabilistic decision function determines the class (spontaneous or posed) for the entire brow action. When tested on 189 samples taken from three different sets of spontaneous and volitional facial data, we attain a 90.7% correct recognition rate.

This paper addresses the problem of human-action recognition by introducing a sparse representation of image sequences as a collection of spatiotemporal events that are localized at points that are salient both in space and time. The spatiotemporal salient points are detected by measuring the variations in the information content of pixel neighborhoods not only in space but also in time. An appropriate distance metric between two collections of spatiotemporal salient points is introduced, which is based on the chamfer distance and an iterative linear time-warping technique that deals with time expansion or time-compression issues. A classification scheme that is based on relevance vector machines and on the proposed distance measure is proposed. Results on real image sequences from a small database depicting people performing 19 aerobic exercises are presented.

Automatic analysis of human facial expression is a challenging problem with many applications. Most of the existing automated systems for facial expression analysis attempt to recognize a few prototypic emotional expressions, such as anger and happiness. Instead of representing another approach to machine analysis of prototypic facial expressions of emotion, the method presented in this paper attempts to handle a large range of human facial behavior by recognizing facial muscle actions that produce expressions. Virtually all of the existing vision systems for facial muscle action detection deal only with frontal-view face images and cannot handle temporal dynamics of facial actions. In this paper, we present a system for automatic recognition of facial action units (AUs) and their temporal models from long, profile-view face image sequences. We exploit particle filtering to track 15 facial points in an input face-profile sequence, and we introduce facial-action-dynamics recognition from continuous video input using temporal rules. The algorithm performs both automatic segmentation of an input video into facial expressions pictured and recognition of temporal segments (i.e., onset, apex, offset) of 27 AUs occurring alone or in a combination in the input face-profile video. A recognition rate of 87% is achieved.

This paper addresses the problem of human action recognition by introducing a sparse representation of image sequences as a collection of spatiotemporal events that are localized at points that are salient both in space and time. We detect the spatiotemporal salient points by measuring changes in the information content of pixel neighborhoods not only in space but also in time. We introduce an appropriate distance metric between two collections of spatiotemporal salient points that is based on the Chamfer distance and an iterative linear time warping technique that deals with time expansion or time compression issues. We propose a classification scheme that is based on relevance vector machines and on the proposed distance measure. We present results on real image sequences from a small database depicting people performing 19 aerobic exercises.

In the last decade, the research topic of automatic analysis of facial expressions has become a central topic in machine vision research. Nonetheless, there is a glaring lack of a comprehensive, readily accessible reference set of face images that could be used as a basis for benchmarks for efforts in the field. This lack of easily accessible, suitable, common testing resource forms the major impediment to comparing and extending the issues concerned with automatic facial expression analysis. In this paper, we discuss a number of issues that make the problem of creating a benchmark facial expression database difficult. We then present the MMI facial expression database, which includes more than 1500 samples of both static images and image sequences of faces in frontal and in profile view displaying various expressions of emotion, single and multiple facial muscle activation. It has been built as a Web-based direct-manipulation application, allowing easy access and easy search of the available images. This database represents the most comprehensive reference set of images for studies on facial expression analysis to date.

Social and emotional intelligence are aspects of human intelligence that have been argued to be better predictors than IQ for measuring aspects of success in life, especially in social interactions, learning, and adapting to what is important. When it comes to machines, not all of them will need such skills. Yet to have machines like computers, broadcast systems, and cars, capable of adapting to their users and of anticipating their wishes, endowing them with the ability to recognize user's affective states is necessary. This article discusses the components of human affect, how they might be integrated into computers, and how far are we from realizing affective multimodal human-computer interaction.

The recognition of facial expressions in image sequences is a difficult problem with many applications in human-machine interaction. Facial expression analyzers achieve good recognition rates, but virtually all of them deal only with prototypic facial expressions of emotions and cannot handle temporal dynamics of facial displays. The method presented here attempts to handle a large range of human facial behavior by recognizing facial action units (AUs) and their temporal segments (i.e., onset, apex, offset) that produce expressions. We exploit particle filtering to track 20 facial points in an input face video and we introduce AU-dynamics recognition using temporal rules. When tested on Cohn-Kanade and MMI facial expression databases, the proposed method achieved a recognition rate of 90% when detecting 27 AUs occurring alone or in a combination in an input face image sequence.

This paper addresses the problem of template-based tracking of non rigid objects. We use the well-known framework of auxiliary particle filtering and propose an observation model that explicitly addresses appearance changes that are caused by local deformations of the tracked object. In addition, by adopting a colour difference that is invariant to local changes in the illumination, the proposed observation model can deal with changing lighting conditions and shadows. Experimental results with real image sequences demonstrate the efficiency of the proposed method in tracking facial features, such as mouth and eye corners

Locating facial feature points in images of faces is an important stage for numerous facial image interpretation tasks. In this paper we present a method for fully automatic detection of 20 facial feature points in images of expressionless faces using Gabor feature based boosted classifiers. The method adopts fast and robust face detection algorithm, which represents an adapted version of the original Viola-Jones face detector. The detected face region is then divided into 20 relevant regions of interest, each of which is examined further to predict the location of the facial feature points. The proposed facial feature point detection method uses individual feature patch templates to detect points in the relevant region of interest. These feature models are GentleBoost templates built from both gray level intensities and Gabor wavelet features. When tested on the Cohn-Kanade database, the method has achieved average recognition rates of 93%.

This work presents an automatic system that we developed for automatic recognition of facial gestures (facial muscle activity) from static images of combined frontal-and profile-view of the face. For the frontal view, the face region is subjected to multi-detector processing which per facial component (eyes, eyebrows, mouth), generates a spatial sample of its contour. A set of 19 frontal-face feature points is then extracted from the spatially sampled contours of the facial features. For the profile view, 10 feature points are extracted from the contour of the face-profile region. Based on these 29 points, 29 individual facial muscle action units (AUs) occurring alone or in combinations in an input dual-view image are recognized using a rule-based reasoning. With each scored AU, the utilized algorithm associates a factor denoting the certainty with which the pertinent AU has been scored. A recognition rate of 86% is achieved.

The recognition of facial action units (AUs) in image sequences is a challenging problem. AU detectors achieve good recognition rates, but virtually all of them deal only with frontal-view face images and cannot handle the temporal dynamics of AUs. We report on a system for automatic recognition of temporal models of AUs from long, profile-view, face image sequences. We exploit particle filtering to track 15 facial points in an input face-profile video sequence and we introduce facial-behavior temporal-dynamics recognition from continuous video input using temporal rules. The utilized algorithm performs both automatic segmentation and recognition of temporal segments (i.e., onset, apex, offset) of 23 AUs occurring alone or in a combination in an input face-profile video sequence. A recognition rate of 88% is achieved.

To allow for rich and sometimes subtle shadings of emotion that humans recognize in a facial expression, user-profiled recognition of emotions from images of faces is needed. In this work, we introduce a case-based reasoning system capable of classifying facial expressions (given in terms of facial muscle actions) into the emotion categories learned from the user. The utilized case base is a dynamic, incrementally self-organizing-event-content-addressable memory that allows fact retrieval and evaluation of encountered events, based upon the user preferences and the generalizations formed from prior input. Two versions of a prototype system are presented: one aims at recognition of six "universal" emotions and the other aims at recognition of affective states learned from the user. Validation studies suggest that in 100% and in 97% of the test cases, respectively, interpretations produced by the system are consistent with those of the two users who trained the two versions of the prototype system.

Automatic recognition of facial gestures (i.e., facial muscle activity) is rapidly becoming an area of intense interest in the research field of machine vision. In this paper, we present an automated system that we developed to recognize facial gestures in static, frontal- and/or profile-view color face images. A multidetector approach to facial feature localization is utilized to spatially sample the profile contour and the contours of the facial components such as the eyes and the mouth. From the extracted contours of the facial features, we extract ten profile-contour fiducial points and 19 fiducial points of the contours of the facial components. Based on these, 32 individual facial muscle actions (AUs) occurring alone or in combination are recognized using rule-based reasoning. With each scored AU, the utilized algorithm associates a factor denoting the certainty with which the pertinent AU has been scored. A recognition rate of 86% is achieved.

The ability to recognize affective states of a person we are communicating with is the core of emotional intelligence. Emotional intelligence is a facet of human intelligence that has been argued to be indispensable and perhaps the most important for successful interpersonal social interaction. This paper argues that next-generation human-computer interaction (HCI) designs need to include the essence of emotional intelligence - the ability to recognize a user's affective states-in order to become more human-like, more effective, and more efficient. Affective arousal modulates all nonverbal communicative cues (facial expressions, body movements, and vocal and physiological reactions). In a face-to-face interaction, humans detect and interpret those interactive signals of their communicator with little or no effort. Yet design and development of an automated system that accomplishes these tasks is rather difficult. This paper surveys the past work in solving these problems by a computer and provides a set of recommendations for developing the first part of an intelligent multimodal HCI-an automatic personalized analyzer of a user's nonverbal affective feedback.

A robust way to discern facial gestures in images of faces, insensitive to scale, pose, and occlusion, is still the key research challenge in the automatic facial-expression analysis domain. A practical method recognized as the most promising one for addressing this problem is through a facial-gesture analysis of multiple views of the face. Yet, current systems for automatic facial-gesture analysis utilize mainly portrait or nearly frontal views of faces. To advance the existing technological framework upon which research on automatic facial-gesture analysis from multiple facial views can be based, we developed an automatic system as to analyze subtle changes in facial expressions based on profile-contour reference points in a profile-view video. A probabilistic classification method based on statistical modeling of the color and motion properties of the profile in the scene is proposed for tracking the profile face. From the segmented profile face, we extract the profile contour and from it, we extract 10 profile-contour reference points. Based on these, 20 individual facial muscle actions occurring alone or in a combination are recognized by a rule-based method. A recognition rate of 85% is achieved.

Automatic analysis of facial gestures is rapidly becoming an area of intense interest in computer science and human-computer interaction design communities. However, the basic goal of this area of research translating detected facial changes into a human-like description of shown, facial expression is yet to be achieved. One of the main impediments to achieving this aim is the fad that human interpretations of a facial expression differ depending upon whether the observed person is speaking or not. A first step in tackling this problem is to achieve automatic detection of facial gestures that are typical for speech articulation. This paper presents our approach to seizing this step in the research on automatic facial expression analysis. It presents: a robust and flexible method for recognition of 22 facial muscle actions from face image sequences, a method for automatic determination of whether the observed subject is speaking or not, and an experimental study on facial muscle actions typical for speech articulation.

This paper presents a novel, robust and flexible method for extracting four mouth features (top of the upper lip, bottom of the lower lip, left and right mouth corners) from facial image sequences. While robustness is referred to subject variability, pose, and image quality, flexibility is begotten by efficient fusion of several information sources and expounding the certainty of the generated results.

The authors attempt to automate facial expression recognition and introduce it into man-machine interaction as a new modality. This will make the interaction compact and more efficient. As the first step, we developed a self-adaptive expert system that accepts the facial feature contours localized in a static dual-view facial image and returns the expression interpretation label(s) used by the user. Expression identification in terms of the encountered facial actions is also displayed to the user. Reasoning with uncertainty about the extracted facial expression data is employed for facial action coding and quantification. A memory of experiences, inspired by R.C. Schank's (1984) theory of human autobiographical memory organization and instance based learning, expounds the encoded facial actions in terms of the learned interpretation labels. Validation studies on the prototype suggest that the expression identification and interpretation achieved are generally consistent with those defined by the users.

Humans detect and interpret faces and facial expressions in a scene with little or no effort. Still, development of an automated system that accomplishes this task is rather difficult. There are several related problems: detection of an image segment as a face, extraction of the facial expression information, and classification of the expression (e.g., in emotion categories). A system that performs these operations accurately and in real time would form a big step in achieving a human-like interaction between man and machine. The paper surveys the past work in solving these problems. The capability of the human visual system with respect to these problems is discussed, too. It is meant to serve as an ultimate goal and a guide for determining recommendations for development of an automatic facial expression analyzer.

This paper discusses the Integrated System for Facial Expression Recognition (ISFER), which performs facial expression analysis from a still dual facial view image. The system consists of three major parts: a facial data generator, a facial data evaluator and a facial data analyser. While the facial data generator applies fairly conventional techniques for facial feature extraction, the rest of the system represents a novel way of performing a reliable identification of 30 different face actions and a multiple classification of expressions into the six basic emotion categories. An expert system has been utilised to convert low level face geometry into high level face actions, and then this into highest level weighted emotion labels. The system evaluation results demonstrated rather high concurrent validity with human coding of facial expressions using FACS and formal instructions in emotion signals.