Data compression is a critical component of today's digital infrastructure, in particular for storage and transmission of visual data over the internet. While end-to-end neural codecs have been shown to outperform hand-designed image and video codecs, their mainstream adoption is hindered by their high computational cost. In this paper we reassess standard design choices for neural image compression autoencoders, identifying two inefficiencies: the first is the substantial compute spent on high-resolution feature maps, and the second is methods of applying activations that lead to less expressive representations. We mitigate both issues and propose a new architecture, PatchMixer, that begins with approximately patchwise-independent encoding, followed by mixing layers, thus enabling compute savings. PatchMixer achieves a BD-rate within 5% of VVC on Kodak with 97 kMACs per pixel of decoding complexity and 200 kMACs per pixel of cumulative complexity, to our knowledge the lowest cumulative complexity for a neural codec within 5% of VVC.

Latent diffusion models (LDMs) power state-of-the-art high-resolution generative image models. LDMs learn the data distribution in the latent space of an autoencoder (AE) and produce images by mapping the generated latents into RGB image space using the AE decoder. While this approach allows for efficient model training and sampling, it induces a disconnect between the training of the diffusion model and the decoder, resulting in a loss of detail in the generated images. To remediate this disconnect, we propose to leverage the internal features of the decoder to define a latent perceptual loss (LPL). This loss encourages the models to create sharper and more realistic images. Our loss can be seamlessly integrated with common autoencoders used in latent diffusion models, and can be applied to different generative modeling paradigms such as DDPM with epsilon and velocity prediction, as well as flow matching. Extensive experiments with models trained on three datasets at 256 and 512 resolution show improved quantitative -- with boosts between 6% and 20% in FID -- and qualitative results when using our perceptual loss.

Vector quantization is a fundamental technique for compression and large-scale nearest neighbor search. For high-accuracy operating points, multi-codebook quantization associates data vectors with one element from each of multiple codebooks. An example is residual quantization (RQ), which iteratively quantizes the residual error of previous steps. Dependencies between the different parts of the code are, however, ignored in RQ, which leads to suboptimal rate-distortion performance. Qinco recently addressed this inefficiency by using a neural network to determine the quantization codebook in RQ based on the vector reconstruction from previous steps. In this paper we introduce Qinco2 which extends and improves Qinco with (i) improved vector encoding using codeword pre-selection and beam-search, (ii) a fast approximate decoder leveraging codeword pairs to establish accurate short-lists for search, and (iii) an optimized training procedure and network architecture. We conduct experiments on four datasets to evaluate Qinco2 for vector compression and billion-scale nearest neighbor search. We obtain outstanding results in both settings, improving the state-of-the-art reconstruction MSE by 44% for 16-byte vector compression on BigANN, and search accuracy by 24% with 8-byte encodings on Deep1M.

Inspired by the principle of deliberate practice in human learning, we propose Deliberate Practice for Synthetic Data Generation (DP), a novel framework that improves sample efficiency through dynamic synthetic data generation. Prior work has shown that scaling synthetic data is inherently challenging, as naively adding new data leads to diminishing returns. To address this, pruning has been identified as a key mechanism for improving scaling, enabling models to focus on the most informative synthetic samples. Rather than generating a large dataset and pruning it afterward, DP efficiently approximates the direct generation of informative samples. We theoretically show how training on challenging, informative examples improves scaling laws and empirically validate that DP achieves better scaling performance with significantly fewer training samples and iterations. On ImageNet-100, DP generates 3.4x fewer samples and requires six times fewer iterations, while on ImageNet-1k, it generates 8x fewer samples with a 30% reduction in iterations, all while achieving superior performance compared to prior work.

Guidance techniques are commonly used in diffusion and flow models to improve image quality and input consistency for conditional generative tasks such as class-conditional and text-to-image generation. In particular, classifier-free guidance (CFG) is the most widely adopted guidance technique. It results, however, in trade-offs across quality, diversity and consistency: improving some at the expense of others. While recent work has shown that it is possible to disentangle these factors to some extent, such methods come with an overhead of requiring an additional (weaker) model, or require more forward passes per sampling step. In this paper, we propose Entropy Rectifying Guidance (ERG), a simple and effective guidance method based on inference-time changes in the attention mechanism of state-of-the-art diffusion transformer architectures, which allows for simultaneous improvements over image quality, diversity and prompt consistency. ERG is more general than CFG and similar guidance techniques, as it extends to unconditional sampling. We show that ERG results in significant improvements in various generation tasks such as text-to-image, class-conditional and unconditional image generation. We also show that ERG can be seamlessly combined with other recent guidance methods such as CADS and APG, further improving generations.

Conditional image generative models hold considerable promise to produce infinite amounts of synthetic training data. Yet, recent progress in generation quality has come at the expense of generation diversity, limiting the utility of these models as a source of synthetic training data. Although guidance-based approaches have been introduced to improve the utility of generated data by focusing on quality or diversity, the (implicit or explicit) utility functions oftentimes disregard the potential distribution shift between synthetic and real data. In this work, we introduce Chamfer Guidance: a training-free guidance approach which leverages a handful of real exemplar images to characterize the quality and diversity of synthetic data. We show that by leveraging the proposed Chamfer Guidance, we can boost the diversity of the generations w.r.t. a dataset of real images while maintaining or improving the generation quality on ImageNet-1k and standard geo-diversity benchmarks. Our approach achieves state-of-the-art few-shot performance with as little as 2 exemplar real images, obtaining 96.4% in terms of precision, and 86.4% in terms of distributional coverage, which increase to 97.5% and 92.7%, respectively, when using 32 real images. We showcase the benefits of the Chamfer Guidance generation by training downstream image classifiers on synthetic data, achieving accuracy boost of up to 15% for in-distribution over the baselines, and up to 16\% in out-of-distribution. Furthermore, our approach does not require using the unconditional model, and thus obtains a 31\% reduction in FLOPs w.r.t. classifier-free-guidance-based approaches at sampling time.

Semantic image synthesis, i.e., generating images from user-provided semantic label maps, is an important conditional image generation task as it allows to control both the content as well as the spatial layout of generated images. Although diffusion models have pushed the state of the art in generative image modeling, the iterative nature of their inference process makes them computationally demanding. Other approaches such as GANs are more efficient as they only need a single feed-forward pass for generation, but the image quality tends to suffer when modeling large and diverse datasets. In this work, we propose a new class of GAN discriminators for semantic image synthesis that generates highly realistic images by exploiting feature backbones pretrained for tasks such as image classification. We also introduce a new generator architecture with better context modeling and using cross-attention to inject noise into latent variables, leading to more diverse generated images. Our model, which we dub DP-SIMS, achieves state-of-the-art results in terms of image quality and consistency with the input label maps on ADE-20K, COCO-Stuff, and Cityscapes, surpassing recent diffusion models while requiring two orders of magnitude less compute for inference.

Image codecs are typically optimized to trade-off bitrate vs. distortion metrics. At low bitrates, this leads to compression artefacts which are easily perceptible, even when training with perceptual or adversarial losses. To improve image quality and remove dependency on the bitrate we propose to decode with iterative diffusion models. We condition the decoding process on a vector-quantized image representation, as well as a global image description to provide additional context. We dub our model `PerCo'' for ``perceptual compression'', and compare it to state-of-the-art codecs at rates from 0.1 down to 0.003 bits per pixel. The latter rate is more than an order of magnitude smaller than those considered in most prior work, compressing a 512x768 Kodak image with less than 153 bytes. Despite this ultra-low bitrate, our approach maintains the ability to reconstruct realistic images. We find that our model leads to reconstructions with state-of-the-art visual quality as measured by FID and KID. As predicted by rate-distortion-perception theory, visual quality is less dependent on the bitrate than previous methods.

Vector quantization is a fundamental operation for data compression and vector search. To obtain high accuracy, multi-codebook methods represent each vector using codewords across several codebooks. Residual quantization (RQ) is one such method, which iteratively quantizes the error of the previous step. While the error distribution is dependent on previously-selected codewords, this dependency is not accounted for in conventional RQ as it uses a fixed codebook per quantization step. In this paper, we propose QINCo, a neural RQ variant that constructs specialized codebooks per step that depend on the approximation of the vector from previous steps. Experiments show that QINCo outperforms state-of-the-art methods by a large margin on several datasets and code sizes. For example, QINCo achieves better nearest-neighbor search accuracy using 12-byte codes than the state-of-the-art UNQ using 16 bytes on the BigANN1M and Deep1M datasets.

Large-scale training of latent diffusion models (LDMs) has enabled unprecedented quality in image generation. However, large-scale end-to-end training of these models is computationally costly, and hence most research focuses either on finetuning pretrained models or experiments at smaller scales. In this work we aim to improve the training efficiency and performance of LDMs with the goal of scaling to larger datasets and higher resolutions. We focus our study on two points that are critical for good performance and efficient training: (i) the mechanisms used for semantic level (\eg a text prompt, or class name) and low-level (crop size, random flip, \etc) conditioning of the model, and (ii) pre-training strategies to transfer representations learned on smaller and lower-resolution datasets to larger ones. The main contributions of our work are the following: we present systematic experimental study of these points, we propose a novel conditioning mechanism that disentangles semantic and low-level conditioning, we obtain state-of-the-art performance on CC12M for text-to-image at 512 resolution.

Although instance segmentation methods have improved considerably, the dominant paradigm is to rely on fully-annotated training images, which are tedious to obtain. To alleviate this reliance, and boost results, semi-supervised approaches leverage unlabeled data as an additional training signal that limits overfitting to the labeled samples. In this context, we present novel design choices to significantly improve teacher-student distillation models. In particular, we (i) improve the distillation approach by introducing a novel "guided burn-in" stage, and (ii) evaluate different instance segmentation architectures, as well as backbone networks and pre-training strategies. Contrary to previous work which uses only supervised data for the burn-in period of the student model, we also use guidance of the teacher model to exploit unlabeled data in the burn-in period. Our improved distillation approach leads to substantial improvements over previous state-of-the-art results. For example, on the Cityscapes dataset we improve mask-AP from 23.7 to 33.9 when using labels for 10% of images, and on the COCO dataset we improve mask-AP from 18.3 to 34.1 when using labels for only 1% of the training data.

Semantic image synthesis aims to generate photo realistic images given a semantic segmentation map. Despite much recent progress, training them still requires large datasets of images annotated with per-pixel label maps that are extremely tedious to obtain. To alleviate the high annotation cost, we propose a transfer method that leverages a model trained on a large source dataset to improve the learning ability on small target datasets via estimated pairwise relations between source and target classes. The class affinity matrix is introduced as a first layer to the source model to make it compatible with the target label maps, and the source model is then further finetuned for the target domain. To estimate the class affinities we consider different approaches to leverage prior knowledge: semantic segmentation on the source domain, textual label embeddings, and self-supervised vision features. We apply our approach to GAN-based and diffusion-based architectures for semantic synthesis. Our experiments show that the different ways to estimate class affinity can be effectively combined, and that our approach significantly improves over existing state-of-the-art transfer approaches for generative image models.

We introduce submodel co-training, a regularization method related to co-training, self-distillation and stochastic depth. Given a neural network to be trained, for each sample we implicitly instantiate two altered networks, “submodels”, with stochastic depth: we activate only a subset of the layers. Each network serves as a soft teacher to the other, by providing a loss that complements the regular loss provided by the one-hot label. Our approach, dubbed “co-sub”, uses a single set of weights, and does not involve a pre-trained external model or temporal averaging. Experimentally, we show that submodel co-training is effective to train backbones for recognition tasks such as image classification and semantic segmentation. Our approach is compatible with multiple architectures, including RegNet, ViT, PiT, XCiT, Swin and ConvNext. Our training strategy improves their results in comparable settings. For instance, a ViT-B pretrained with cosub on ImageNet-21k obtains 87.4% top1 acc. @448 on ImageNet-val.

Large-scale text-to-image diffusion models have significantly improved the state of the art in generative image modeling and allow for an intuitive and powerful user interface to drive the image generation process. Expressing spatial constraints, e.g. to position specific objects in particular locations, is cumbersome using text; and current text-based image generation models are not able to accurately follow such instructions. In this paper we consider image generation from text associated with segments on the image canvas, which combines an intuitive natural language interface with precise spatial control over the generated content. We propose ZestGuide, a "zero-shot" segmentation guidance approach that can be plugged into pre-trained text-to-image diffusion models, and does not require any additional training. It leverages implicit segmentation maps that can be extracted from cross-attention layers, and uses them to align the generation with input masks. Our experimental results combine high image quality with accurate alignment of generated content with input segmentations, and improve over prior work both quantitatively and qualitatively, including methods that require training on images with corresponding segmentations. Compared to Paint with Words, the previous state-of-the art in image generation with zero-shot segmentation conditioning, we improve by 5 to 10 mIoU points on the COCO dataset with similar FID scores.

Lossy image compression aims to represent images in as few bits as possible while maintaining fidelity to the original. Theoretical results indicate that optimizing distortion metrics such as PSNR or MS-SSIM necessarily leads to a discrepancy in the statistics of original images from those of reconstructions, in particular at low bitrates, often manifested by the blurring of the compressed images. Previous work has leveraged adversarial discriminators to improve statistical fidelity. Yet these binary discriminators adopted from generative modeling tasks may not be ideal for image compression. In this paper, we introduce a non-binary discriminator that is conditioned on quantized local image representations obtained via VQ-VAE autoencoders. Our evaluations on the CLIC2020, DIV2K and Kodak datasets show that our discriminator is more effective for jointly optimizing distortion (e.g., PSNR) and statistical fidelity (e.g., FID) than the PatchGAN of the state-of-the-art HiFiC model. On CLIC2020, we obtain the same FID as HiFiC with 30-40% fewer bits.

We present ResMLP, an architecture built entirely upon multi-layer perceptrons for image classification. It is a simple residual network that alternates (i) a linear layer in which image patches interact, independently and identically across channels, and (ii) a two-layer feed-forward network in which channels interact independently per patch. When trained with a modern training strategy using heavy data-augmentation and optionally distillation, it attains surprisingly good accuracy/complexity trade-offs on ImageNet. We also train ResMLP models in a self-supervised setup, to further remove priors from employing a labelled dataset. Finally, by adapting our model to machine translation we achieve surprisingly good results. We share pre-trained models and our code based on the Timm library.

Deep generative models, like GANs, have considerably improved the state of the art in image synthesis, and are able to generate near photo-realistic images in structured domains such as human faces. Based on this success, recent work on image editing proceeds by projecting images to the GAN latent space and manipulating the latent vector. However, these approaches are limited in that only images from a narrow domain can be transformed, and with only a limited number of editing operations. We propose FlexIT, a novel method which can take any input image and a user-defined text instruction for editing. Our method achieves flexible and natural editing, pushing the limits of semantic image translation. First, FlexIT combines the input image and text into a single target point in the CLIP multimodal embedding space. Via the latent space of an autoencoder, we iteratively transform the input image toward the target point, ensuring coherence and quality with a variety of novel regularization terms. We propose an evaluation protocol for semantic image translation, and thoroughly evaluate our method on ImageNet. Code will be available at https://github.com/facebookresearch/SemanticImageTranslation/.

Convolutional networks have been extremely successful for regular data structures such as 2D images and 3D voxel grids. The transposition to meshes is, however, not straightforward due to their irregular structure. We explore how the dual, face-based representation of triangular meshes can be leveraged as a data structure for graph convolutional networks. In the dual mesh, each node (face) has a fixed number of neighbors, which makes the networks less susceptible to overfitting on the mesh topology, and also allows the use of input features that are naturally defined over faces, such as surface normals and face areas. We evaluate the dual approach on the shape correspondence task on the Faust human shape dataset and variants of it with different mesh topologies. Our experiments show that results of graph convolutional networks improve when defined over the dual rather than primal mesh. Moreover, our models that explicitly leverage the neighborhood regularity of dual meshes allow improving results further while being more robust to changes in the mesh topology.

Inference in deep neural networks can be computationally expensive, and networks capable of anytime inference are important in scenarios where the amount of compute or input data varies over time. In such networks the inference process can interrupted to provide a result faster, or continued to obtain a more accurate result. We propose Hierarchical Neural Ensembles (HNE), a novel framework to embed an ensemble of multiple networks in a hierarchical tree structure, sharing intermediate layers. In HNE we control the complexity of inference on-the-fly by evaluating more or less models in the ensemble. Our second contribution is a novel hierarchical distillation method to boost the predictions of small ensembles. This approach leverages the nested structure of our ensembles, to optimally allocate accuracy and diversity across the individual models. Our experiments show that, compared to previous anytime inference models, HNE provides state-of-the-art accuracy-computation trade-offs on the CIFAR-10/100 and ImageNet datasets.

Following their success in natural language processing, transformers have recently shown much promise for computer vision. The self-attention operation underlying transformers yields global interactions between all tokens ,i.e. words or image patches, and enables flexible modelling of image data beyond the local interactions of convolutions. This flexibility, however, comes with a quadratic complexity in time and memory, hindering application to long sequences and high-resolution images. We propose a “transposed” version of self-attention that operates across feature channels rather than tokens, where the interactions are based on the cross-covariance matrix between keys and queries. The resulting cross-covariance attention (XCA) has linear complexity in the number of tokens, and allows efficient processing of high-resolution images.Our cross-covariance image transformer (XCiT) is built upon XCA. It combines the accuracy of conventional transformers with the scalability of convolutional architectures. We validate the effectiveness and generality of XCiT by reporting excellent results on multiple vision benchmarks, including image classification and self-supervised feature learning on ImageNet-1k, object detection and instance segmentation on COCO, and semantic segmentation on ADE20k.We will opensource our code and trained models to reproduce the reported results.

Generative Adversarial Networks (GANs) can generate near photo realistic images in narrow domains such as human faces. Yet, modeling complex distributions of datasets such as ImageNet and COCO-Stuff remains challenging in unconditional settings. In this paper, we take inspiration from kernel density estimation techniques and introduce a non-parametric approach to modeling distributions of complex datasets. We partition the data manifold into a mixture of overlapping neighborhoods described by a datapoint and its nearest neighbors, and introduce a model, called instance-conditioned GAN (IC-GAN), which learns the distribution around each datapoint. Experimental results on ImageNet and COCO-Stuff show that IC-GAN significantly improves over unconditional models and unsupervised data partitioning baselines. Moreover, we show that IC-GAN can effortlessly transfer to datasets not seen during training by simply changing the conditioning instances, and still generate realistic images. Finally, we extend IC-GAN to the class-conditional case and show semantically controllable generation and competitive quantitative results on ImageNet; while improving over BigGAN on ImageNet-LT. Code and trained models to reproduce the reported results are available at https://github.com/facebookresearch/ic_gan.

Maha Elbayad, Michael Ustaszewski, Emmanuelle Esperança-Rodier, Francis Brunet-Manquat, Jakob Verbeek, Laurent Besacier. Proceedings of the 28th International Conference on Computational Linguistics. 2020.

Visual localization is critical to many applications in computer vision and robotics. To address single-image RGB localization, state-of-the-art feature-based methods match local descriptors between a query image and a pre-built 3D model. Recently, deep neural networks have been exploited to regress the mapping between raw pixels and 3D coordinates in the scene, and thus the matching is implicitly performed by the forward pass through the network. However, in a large and ambiguous environment, learning such a regression task directly can be difficult for a single network. In this work, we present a new hierarchical scene coordinate network to predict pixel scene coordinates in a coarse-to-fine manner from a single RGB image. The network consists of a series of output layers, each of them conditioned on the previous ones. The final output layer predicts the 3D coordinates and the others produce progressively finer discrete location labels. The proposed method outperforms the baseline regression-only network and allows us to train compact models which scale robustly to large environments. It sets a new state-of-the-art for single-image RGB localization performance on the 7-Scenes, 12-Scenes, Cambridge Landmarks datasets, and three combined scenes. Moreover, for large-scale outdoor localization on the Aachen Day-Night dataset, we present a hybrid approach which outperforms existing scene coordinate regression methods, and reduces significantly the performance gap w.r.t. explicit feature matching methods.

We propose a novel amortized variational inference scheme for an empirical Bayes meta-learning model, where model parameters are treated as latent variables. We learn the prior distribution over model parameters conditioned on limited training data using a variational autoencoder approach. Our framework proposes sharing the same amortized inference network between the conditional prior and variational posterior distributions over the model parameters. While the posterior leverages both the labeled support and query data, the conditional prior is based only on the labeled support data. We show that in earlier work, relying on Monte-Carlo approximation, the conditional prior collapses to a Dirac delta function. In contrast, our variational approach prevents this collapse and preserves uncertainty over the model parameters. We evaluate our approach on the miniImageNet, CIFAR-FS and FC100 datasets, and present results demonstrating its advantages over previous work.

Simultaneous machine translation consists in starting output generation before the entire input sequence is available. Wait-k decoders offer a simple but efficient approach for this problem. They first read k source tokens, after which they alternate between producing a target token and reading another source token. We investigate the behavior of wait-k decoding in low resource settings for spoken corpora using IWSLT datasets. We improve training of these models using unidirectional encoders, and training across multiple values of k. Experiments with Transformer and 2D-convolutional architectures show that our wait-k models generalize well across a wide range of latency levels. We also show that the 2D-convolution architecture is competitive with Transformers for simultaneous translation of spoken language.

Despite the outstanding performance of convolutional neural networks (CNNs) for many vision tasks, the required computational cost during inference is problematic when resources are limited. In this context, we propose Convolutional Neural Mixture Models (CNMMs), a probabilistic model embedding a large number of CNNs that can be jointly trained and evaluated in an efficient manner. Within the proposed framework, we present different mechanisms to prune subsets of CNNs from the mixture, allowing to easily adapt the computational cost required for inference. Image classification and semantic segmentation experiments show that our method achieve excellent accuracy-compute trade-offs. Moreover, unlike most of previous approaches, a single CNMM provides a large range of operating points along this trade-off, without any re-training.

We investigate deep Bayesian neural networks with Gaussian priors on the weights and a class of ReLU-like nonlinearities. Bayesian neural networks with Gaussian priors are well known to induce an L2, “weight decay”, regularization. Our results indicate a more intricate regularization effect at the level of the unit activations. Our main result establishes that the induced prior distribution on the units before and after activation becomes increasingly heavy-tailed with the depth of the layer. We show that first layer units are Gaussian, second layer units are sub-exponential, and units in deeper layers are characterized by sub-Weibull distributions. Our results provide new theoretical insight on deep Bayesian neural networks, which we corroborate with simulation experiments.

Unsupervised learning of generative models has seen tremendous progress over recent years, in particular due to generative adversarial networks (GANs), variational autoencoders, and flow-based models. GANs have dramatically improved sample quality, but suffer from two drawbacks: (i) they mode-drop, \ie, do not cover the full support of the train data, and (ii) they do not allow for likelihood evaluations on held-out data. In contrast likelihood-based training encourages models to cover the full support of the train data, but yields poorer samples. These mutual shortcomings can in principle be addressed by training generative latent variable models in a hybrid adversarial-likelihood manner. However, we show that commonly made parametric assumptions create a conflict between them, making successful hybrid models non trivial. As a solution, we propose the use of deep invertible transformations in the latent variable decoder. This approach allows for likelihood computations in image space, is more efficient than fully invertible models, and can take full advantage of adversarial training. We show that our model significantly improves over existing hybrid models: offering GAN-like samples, IS and FID scores that are competitive with fully adversarial models and improved likelihood scores.

Despite the effectiveness of recurrent neural network language models, their maximum likelihood estimation suffers from two limitations.It treats all sentences that do not match the ground truth as equally poor, ignoring the structure of the output space.Second, it suffers from "exposure bias": during training tokens are predicted given ground-truth sequences, while at test time prediction is conditioned on generated output sequences.To overcome these limitations we build upon the recent reward augmented maximum likelihood approach i.e. sequence-level smoothing that encourages the model to predict sentences close to the ground truth according to a given performance metric.We extend this approach to token-level loss smoothing, and propose improvements to the sequence-level smoothing approach.Our experiments on two different tasks, image captioning and machine translation, show that token-level and sequence-level loss smoothing are complementary, and significantly improve results.

Current state-of-the-art machine translation systems are based on encoder-decoder architectures, that first encode the input sequence, and then generate an output sequence based on the input encoding.Both are interfaced with an attention mechanism that recombines a fixed encoding of the source tokens based on the decoder state.We propose an alternative approach which instead relies on a single 2D convolutional neural network across both sequences.Each layer of our network recodes source tokens on the basis of the output sequence produced so far.Attention-like properties are therefore pervasive throughout the network.Our model yields excellent results, outperforming state-of-the-art encoderdecoder systems, while being conceptually simpler and having fewer parameters.

Convolutional neural networks (CNNs) have massively impacted visual recognition in 2D images, and are now ubiquitous in state-of-the-art approaches. CNNs do not easily extend, however, to data that are not represented by regular grids, such as 3D shape meshes or other graph-structured data, to which traditional local convolution operators do not directly apply. To address this problem, we propose a novel graph-convolution operator to establish correspondences between filter weights and graph neighborhoods with arbitrary connectivity. The key novelty of our approach is that these correspondences are dynamically computed from features learned by the network, rather than relying on predefined static coordinates over the graph as in previous work. We obtain excellent experimental results that significantly improve over previous state-of-the-art shape correspondence results. This shows that our approach can learn effective shape representations from raw input coordinates, without relying on shape descriptors.

Generative adversarial networks (GANs) are pow- erful generative models based on providing feed- back to a generative network via a discriminator network. However, the discriminator usually as- sesses individual samples. This prevents the dis- criminator from accessing global distributional statistics of generated samples, and often leads to mode dropping: the generator models only part of the target distribution. We propose to feed the discriminator with mixed batches of true and fake samples, and train it to predict the ratio of true samples in the batch. The latter score does not depend on the order of samples in a batch. Rather than learning this invariance, we introduce a generic permutation-invariant discriminator ar- chitecture. This architecture is provably a uni- versal approximator of all symmetric functions. Experimentally, our approach reduces mode col- lapse in GANs on two synthetic datasets, and obtains good results on the CIFAR10 and CelebA datasets, both qualitatively and quantitatively.

Deep convolutional neural networks have recently proven extremely effective for difficult face recognition problems in uncontrolled settings. To train such networks, very large training sets are needed with millions of labeled images. For some applications, such as near-infrared (NIR) face recognition, such large training data sets are not publicly available and difficult to collect. In this paper, we propose a method to generate very large training data sets of synthetic images by compositing real face images in a given data set. We show that this method enables to learn models from as few as 10 000 training images, which perform on par with models trained from 500 000 images. Using our approach, we also obtain state-of-the-art results on the CASIA NIR-VIS2.0 heterogeneous face recognition data set.

The ability to predict and therefore to anticipate the future is an important attribute of intelligence. It is also of utmost importance in real-time systems, e.g. in robotics or autonomous driving, which depend on visual scene understanding for decision making. While prediction of the raw RGB pixel values in future video frames has been studied in previous work, here we introduce the novel task of predicting semantic segmentations of future frames. Given a sequence of video frames, our goal is to predict segmentation maps of not yet observed video frames that lie up to a second or further in the future. We develop an autoregressive convolutional neural network that learns to iteratively generate multiple frames. Our results on the Cityscapes dataset show that directly predicting future segmentations is substantially better than predicting and then segmenting future RGB frames. Prediction results up to half a second in the future are visually convincing and are much more accurate than those of a baseline based on warping semantic segmentations using optical flow.

We propose “Areas of Attention”, a novel attentionbased model for automatic image captioning. Our approach models the dependencies between image regions, caption words, and the state of an RNN language model, using three pairwise interactions. In contrast to previous attentionbased approaches that associate image regions only to the RNN state, our method allows a direct association between caption words and image regions. During training these associations are inferred from image-level captions, akin to weakly-supervised object detector training. These associations help to improve captioning by localizing the corresponding regions during testing. We also propose and compare different ways of generating attention areas: CNN activation grids, object proposals, and spatial transformers nets applied in a convolutional fashion. Spatial transformers give the best results. They allow for image specific attention areas, and can be trained jointly with the rest of the network. Our attention mechanism and spatial transformer attention areas together yield state-of-the-art results on the MSCOCO dataset.

Object category localization is a challenging problem in computer vision. Standard supervised training requires bounding box annotations of object instances. This time-consuming annotation process is sidestepped in weakly supervised learning. In this case, the supervised information is restricted to binary labels that indicate the absence/presence of object instances in the image, without their locations. We follow a multiple-instance learning approach that iteratively trains the detector and infers the object locations in the positive training images. Our main contribution is a multi-fold multiple instance learning procedure, which prevents training from prematurely locking onto erroneous object locations. This procedure is particularly important when using high-dimensional representations, such as Fisher vectors and convolutional neural network features. We also propose a window refinement method, which improves the localization accuracy by incorporating an objectness prior. We present a detailed experimental evaluation using the PASCAL VOC 2007 dataset, which verifies the effectiveness of our approach.

Despite the success of CNNs, selecting the optimal architecture for a given task remains an open problem. Instead of aiming to select a single optimal architecture, we propose a ``fabric'' that embeds an exponentially large number of architectures. The fabric consists of a 3D trellis that connects response maps at different layers, scales, and channels with a sparse homogeneous local connectivity pattern. The only hyper-parameters of a fabric are the number of channels and layers. While individual architectures can be recovered as paths, the fabric can in addition ensemble all embedded architectures together, sharing their weights where their paths overlap. Parameters can be learned using standard methods based on back-propagation, at a cost that scales linearly in the fabric size. We present benchmark results competitive with the state of the art for image classification on MNIST and CIFAR10, and for semantic segmentation on the Part Labels dataset.

The bag-of-words (BoW) model treats images as sets of local descriptors and represents them by visual word histograms. The Fisher vector (FV) representation extends BoW, by considering the first and second order statistics of local descriptors. In both representations local descriptors are assumed to be identically and independently distributed (iid), which is a poor assumption from a modeling perspective. It has been experimentally observed that the performance of BoW and FV representations can be improved by employing discounting transformations such as power normalization. In this paper, we introduce non-iid models by treating the model parameters as latent variables which are integrated out, rendering all local regions dependent. Using the Fisher kernel principle we encode an image by the gradient of the data log-likelihood w.r.t. the model hyper-parameters. Our models naturally generate discounting effects in the representations; suggesting that such transformations have proven successful because they closely correspond to the representations obtained for non-iid models. To enable tractable computation, we rely on variational free-energy bounds to learn the hyper-parameters and to compute approximate Fisher kernels. Our experimental evaluation results validate that our models lead to performance improvements comparable to using power normalization, as employed in state-of-the-art feature aggregation methods.

Object category localization is a challenging problem in computer vision. Standard supervised training requires bounding box annotations of object instances. This time-consuming annotation process is sidestepped in weakly supervised learning. In this case, the supervised information is restricted to binary labels that indicate the absence/presence of object instances in the image, without their locations. We follow a multiple-instance learning approach that iteratively trains the detector and infers the object locations in the positive training images. Our main contribution is a multi-fold multiple instance learning procedure, which prevents training from prematurely locking onto erroneous object locations. This procedure is particularly important when high-dimensional representations, such as the Fisher vectors, are used. We present a detailed experimental evaluation using the PASCAL VOC 2007 dataset. Compared to state-of-the-art weakly supervised detectors, our approach better localizes objects in the training images, which translates into improved detection performance.

The Fisher vector (FV) representation is a high-dimensional extension of the popular bag-of-word representation. Transformation of the FV by power and ℓ2normalizations has shown to significantly improve its performance, and led to state-of-the-art results for a range of image and video classification and retrieval tasks. These normalizations, however, render the representation non-additive over local descriptors. Combined with its high dimensionality, this makes the FV computationally expensive for the purpose of localization tasks. In this paper we present approximations to both these normalizations, which yield significant improvements in the memory and computational costs of the FV when used for localization. Second, we show how these approximations can be used to define upper-bounds on the score function that can be efficiently evaluated, which enables the use of branch-and-bound search as an alternative to exhaustive sliding window search. We present experimental evaluation results on classification and temporal localization of actions in videos. These show that the our approximations lead to a speedup of at least one order of magnitude, while maintaining state-of-the-art action recognition and localization performance.

We present an object detection system based on the Fisher vector (FV) image representation computed over SIFT and color descriptors. For computational and storage efficiency, we use a recent segmentation-based method to generate class-independent object detection hypotheses, in combination with data compression techniques. Our main contribution is a method to produce tentative object segmentation masks to suppress background clutter in the features. Re-weighting the local image features based on these masks is shown to improve object detection significantly. We also exploit contextual features in the form of a full-image FV descriptor, and an inter-category rescoring mechanism. Our experiments on the VOC 2007 and 2010 datasets show that our detector improves over the current state-of-the-art detection results.

Action recognition in uncontrolled video is an important and challenging computer vision problem. Recent progress in this area is due to new local features and models that capture spatio-temporal structure between local features, or human-object interactions. Instead of working towards more complex models, we focus on the low-level features and their encoding. We evaluate the use of Fisher vectors as an alternative to bag-of-word histograms to aggregate a small set of state-of-the-art low-level descriptors, in combination with linear classifiers. We present a large and varied set of evaluations, considering (i) classification of short actions in five datasets, (ii) localization of such actions in feature-length movies, and (iii) large-scale recognition of complex events. We find that for basic action recognition and localization MBH features alone are enough for state-of-the-art performance. For complex events we find that SIFT and MFCC features provide complementary cues. On all three problems we obtain state-of-the-art results, while using fewer features and less complex models.

We propose structured prediction models for image labeling that explicitly take into account dependencies among image labels. In our tree-structured models, image labels are nodes, and edges encode dependency relations. To allow for more complex dependencies, we combine labels in a single node and use mixtures of trees. Our models are more expressive than independent predictors, and lead to more accurate label predictions. The gain becomes more significant in an interactive scenario where a user provides the value of some of the image labels at test time. Such an interactive scenario offers an interesting tradeoff between label accuracy and manual labeling effort. The structured models are used to decide which labels should be set by the user, and transfer the user input to more accurate predictions on other image labels. We also apply our models to attribute-based image classification, where attribute predictions of a test image are mapped to class probabilities by means of a given attribute-class mapping. Experimental results on three publicly available benchmark datasets show that in all scenarios our structured models lead to more accurate predictions, and leverage user input much more effectively than state-of-the-art independent models.

We study large-scale image classification methods that can incorporate new classes and training images continuously over time at negligible cost. To this end, we consider two distance-based classifiers, the k-nearest neighbor (k-NN) and nearest class mean (NCM) classifiers, and introduce a new metric learning approach for the latter. We also introduce an extension of the NCM classifier to allow for richer class representations. Experiments on the ImageNet 2010 challenge dataset, which contains over 10(6) training images of 1,000 classes, show that, surprisingly, the NCM classifier compares favorably to the more flexible k-NN classifier. Moreover, the NCM performance is comparable to that of linear SVMs which obtain current state-of-the-art performance. Experimentally, we study the generalization performance to classes that were not used to learn the metrics. Using a metric learned on 1,000 classes, we show results for the ImageNet-10K dataset which contains 10,000 classes, and obtain performance that is competitive with the current state-of-the-art while being orders of magnitude faster. Furthermore, we show how a zero-shot class prior based on the ImageNet hierarchy can improve performance when few training images are available.

The bag-of-words (BoW) model treats images as an unordered set of local regions and represents them by visual word histograms. Implicitly, regions are assumed to be identically and independently distributed (iid), which is a poor assumption from a modeling perspective. We introduce non-iid models by treating the parameters of BoW models as latent variables which are integrated out, rendering all local regions dependent. Using the Fisher kernel we encode an image by the gradient of the data log-likelihood w.r.t. hyper-parameters that control priors on the model parameters. Our representation naturally involves discounting transformations similar to taking square-roots, providing an explanation of why such transformations have proven successful. Using variational inference we extend the basic model to include Gaussian mixtures over local descriptors, and latent topic models to capture the co-occurrence structure of visual words, both improving performance. Our models yield state-of-the-art categorization performance using linear classifiers; without using non-linear transformations such as taking square-roots of features, or using (approximate) explicit embeddings of non-linear kernels.

International audience

We propose structured models for image labeling that take into account the dependencies among the image labels explicitly. These models are more expressive than independent label predictors, and lead to more accurate predictions. While the improvement is modest for fully-automatic image labeling, the gain is significant in an interactive scenario where a user provides the value of some of the image labels. Such an interactive scenario offers an interesting trade-off between accuracy and manual labeling effort. The structured models are used to decide which labels should be set by the user, and transfer the user input to more accurate predictions on other image labels. We also apply our models to attribute-based image classification, where attribute predictions of a test image are mapped to class probabilities by means of a given attribute-class mapping. In this case the structured models are built at the attribute level. We also consider an interactive system where the system asks a user to set some of the attribute values in order to maximally improve class prediction performance. Experimental results on three publicly available benchmark data sets show that in all scenarios our structured models lead to more accurate predictions, and leverage user input much more effectively than state-of-the-art independent models.

The goal of face identification is to decide whether two faces depict the same person or not. This paper addresses the identification problem for face-tracks that are automatically collected from uncontrolled TV video data. Face-track identification is an important component in systems that automatically label characters in TV series or movies based on subtitles and/or scripts: it enables effective transfer of the sparse text-based supervision to other faces. We show that, without manually labeling any examples, metric learning can be effectively used to address this problem. This is possible by using pairs of faces within a track as positive examples, while negative training examples can be generated from pairs of face tracks of different people that appear together in a video frame. In this manner we can learn a cast-specific metric, adapted to the people appearing in a particular video, without using any supervision. Identification performance can be further improved using semi-supervised learning where we also include labels for some of the face tracks. We show that our cast-specific metrics not only improve identification, but also recognition and clustering.

We introduce an extension of bag-of-words image representations to encode spatial layout. Using the Fisher kernel framework we derive a representation that encodes the spatial mean and the variance of image regions associated with visual words. We extend this representation by using a Gaussian mixture model to encode spatial layout, and show that this model is related to a soft-assign version of the spatial pyramid representation. We also combine our representation of spatial layout with the use of Fisher kernels to encode the appearance of local features. Through an extensive experimental evaluation, we show that our representation yields state-of-the-art image categorization results, while being more compact than spatial pyramid representations. In particular, using Fisher kernels to encode both appearance and spatial layout results in an image representation that is computationally efficient, compact, and yields excellent performance while using linear classifiers.

Automatic image annotation is an important tool for keyword-based image retrieval, providing a textual index for non-annotated images. Many image auto annotation methods are based on visual similarity between images to be annotated and images in a training corpus. The annotations of the most similar training images are transferred to the image to be annotated. In this paper we consider using also similarities among the training images, both visual and textual, to derive pseudo relevance models, as well as crossmedia relevance models. We extend a recent state-of-the-art image annotation model to incorporate this information. On two widely used datasets (COREL and IAPR) we show experimentally that the pseudo-relevance models improve the annotation accuracy.

In image categorization the goal is to decide if an image belongs to a certain category or not. A binary classifier can be learned from manually labeled images; while using more labeled examples improves performance, obtaining the image labels is a time consuming process. We are interested in how other sources of information can aid the learning process given a fixed amount of labeled images. In particular, we consider a scenario where keywords are associated with the training images, e.g. as found on photo sharing websites. The goal is to learn a classifier for images alone, but we will use the keywords associated with labeled and unlabeled images to improve the classifier using semi-supervised learning. We first learn a strong Multiple Kernel Learning (MKL) classifier using both the image content and keywords, and use it to score unlabeled images. We then learn classifiers on visual features only, either support vector machines (SVM) or least-squares regression (LSR), from the MKL output values on both the labeled and unlabeled images. In our experiments on 20 classes from the PASCAL VOC'07 set and 38 from the MIR Flickr set, we demonstrate the benefit of our semi-supervised approach over only using the labeled images. We also present results for a scenario where we do not use any manual labeling but directly learn classifiers from the image tags. The semi-supervised approach also improves classification accuracy in this case.

Web image search using text queries has received considerable attention. However, current state-of-the-art approaches require training models for every new query, and are therefore unsuitable for real-world web search applications. The key contribution of this paper is to introduce generic classifiers that are based on query-relative features which can be used for new queries without additional training. They combine textual features, based on the occurence of query terms in web pages and image meta-data, and visual histogram representations of images. The second contribution of the paper is a new database for the evaluation of web image search algorithms. It includes 71478 images returned by a web search engine for 353 different search queries, along with their meta-data and ground-truth annotations. Using this data set, we compared the image ranking performance of our model with that of the search engine, and with an approach that learns a separate classifier for each query. Our generic models that use query-relative features improve significantly over the raw search engine ranking, and also outperform the query-specific models.

We address the problem of continuous stochastic optimal control in the presence of hard obstacles. Due to the non-smooth character of the obstacles, the traditional approach using dynamic programming in combination with function approximation tends to fail. We consider a recently introduced special class of control problems for which the optimal control computation is reformulated in terms of a path integral. The path integral is typically intractable, but amenable to techniques developed for approximate inference. We argue that the variational approach fails in this case due to the non-smooth cost function. Sampling techniques are simple to implement and converge to the exact results given enough samples. However, the infinite cost associated with hard obstacles renders the sampling procedures inefficient in practice. We suggest Expectation Propagation (EP) as a suitable approximation method, and compare the quality and efficiency of the resulting control with an MC sampler on a car steering task and a ball throwing task. We conclude that EP can solve these challenging problems much better than a sampling approach.

This paper introduces the contextual dissimilarity measure, which significantly improves the accuracy of bag-of-features-based image search. Our measure takes into account the local distribution of the vectors and iteratively estimates distance update terms in the spirit of Sinkhorn's scaling algorithm, thereby modifying the neighborhood structure. Experimental results show that our approach gives significantly better results than a standard distance and outperforms the state of the art in terms of accuracy on the Nistér-Stewénius and Lola data sets. This paper also evaluates the impact of a large number of parameters, including the number of descriptors, the clustering method, the visual vocabulary size, and the distance measure. The optimal parameter choice is shown to be quite context-dependent. In particular, using a large number of descriptors is interesting only when using our dissimilarity measure. We have also evaluated two novel variants: multiple assignment and rank aggregation. They are shown to further improve accuracy at the cost of higher memory usage and lower efficiency.

International audience

Image auto-annotation is an important open problem in computer vision. For this task we propose TagProp, a discriminatively trained nearest neighbor model. Tags of test images are predicted using a weighted nearest-neighbor model to exploit labeled training images. Neighbor weights are based on neighbor rank or distance. TagProp allows the integration of metric learning by directly maximizing the log-likelihood of the tag predictions in the training set. In this manner, we can optimally combine a collection of image similarity metrics that cover different aspects of image content, such as local shape descriptors, or global color histograms. We also introduce a word specific sigmoidal modulation of the weighted neighbor tag predictions to boost the recall of rare words. We investigate the performance of different variants of our model and compare to existing work. We present experimental results for three challenging data sets. On all three, TagProp makes a marked improvement as compared to the current state-of-the-art.

Face identification is the problem of determining whether two face images depict the same person or not. This is difficult due to variations in scale, pose, lighting, background, expression, hairstyle, and glasses. In this paper we present two methods for learning robust distance measures: (a) a logistic discriminant approach which learns the metric from a set of labelled image pairs (LDML) and (b) a nearest neighbour approach which computes the probability for two images to belong to the same class (MkNN). We evaluate our approaches on the Labeled Faces in the Wild data set, a large and very challenging data set of faces from Yahoo! News. The evaluation protocol for this data set defines a restricted setting, where a fixed set of positive and negative image pairs is given, as well as an unrestricted one, where faces are labelled by their identity. We are the first to present results for the unrestricted setting, and show that our methods benefit from this richer training data, much more so than the current state-of-the-art method. Our results of 79.3% and 87.5% correct for the restricted and unrestricted setting respectively, significantly improve over the current state-of-the-art result of 78.5%. Confidence scores obtained for face identification can be used for many applications e.g. clustering or recognition from a single training example. We show that our learned metrics also improve performance for these tasks.

Color names are required in real-world applications such as image retrieval and image annotation. Traditionally, they are learned from a collection of labeled color chips. These color chips are labeled with color names within a well-defined experimental setup by human test subjects. However, naming colors in real-world images differs significantly from this experimental setting. In this paper, we investigate how color names learned from color chips compare to color names learned from real-world images. To avoid hand labeling real-world images with color names, we use Google Image to collect a data set. Due to the limitations of Google Image, this data set contains a substantial quantity of wrongly labeled data. We propose several variants of the PLSA model to learn color names from this noisy data. Experimental results show that color names learned from real-world images significantly outperform color names learned from labeled color chips for both image retrieval and image annotation.

We consider two scenarios of naming people in databases of news photos with captions: (i) finding faces of a single person, and (ii) assigning names to all faces. We combine an initial text-based step, that restricts the name assigned to a face to the set of names appearing in the caption, with a second step that analyzes visual features of faces. By searching for groups of highly similar faces that can be associated with a name, the results of purely text-based search can be greatly ameliorated. We improve a recent graph-based approach, in which nodes correspond to faces and edges connect highly similar faces. We introduce constraints when optimizing the objective function, and propose improvements in the low-level methods used to construct the graphs. Furthermore, we generalize the graph-based approach to face naming in the full data set. In this multi-person naming case the optimization quickly becomes computationally demanding, and we present an important speed-up using graph-flows to compute the optimal name assignments in documents. Generative models have previously been proposed to solve the multi-person naming task. We compare the generative and graph-based methods in both scenarios, and find significantly better performance using the graph-based methods in both cases.

Considerable advances have been made in learning to recognize and localize visual object classes. Simple bag-of-feature approaches label each pixel or patch independently. More advanced models attempt to improve the coherence of the labellings by introducing some form of inter-patch coupling: traditional spatial models such as MRF's provide crisper local labellings by exploiting neighbourhood-level couplings, while aspect models such as PLSA and LDA use global relevance estimates (global mixing proportions for the classes appearing in the image) to shape the local choices. We point out that the two approaches are complementary, combining them to produce aspect-based spatial field models that outperform both approaches. We study two spatial models: one based on averaging over forests of minimal spanning trees linking neighboring image regions, the other on an efficient chain-based Expectation Propagation method for regular 8-neighbor Markov random fields. The models can be trained using either patch-level labels or image-level keywords. As input features they use factored observation models combining texture, color and position cues. Experimental results on the MSR Cambridge data sets show that combining spatial and aspect models significantly improves the region-level classification accuracy. In fact our models trained with image-level labels outperform PLSA trained with pixel-level ones.

Within a computer vision context color naming is the action of assigning linguistic color labels to image pixels. In general, research on color naming applies the following paradigm: a collection of color chips is labelled with color names within a well-defined experimental setup by multiple test subjects. The collected data set is subsequently used to label RGB values in real-world images with a color name. Apart from the fact that this collection process is time consuming, it is unclear to what extent color naming within a controlled setup is representative for color naming in real-world images. Therefore we propose to learn color names from real-world images. Furthermore, we avoid test subjects by using Google Image to collect a data set. Due to limitations of Google Image this data set contains a substantial quantity of wrongly labelled data. The color names are learned using a PLSA model adapted to this task. Experimental results show that color names learned from real-world images significantly outperform color names learned from labelled color chips on retrieval and classification.

We propose to use high-level visual information to improve illuminant estimation. Several illuminant estimation approaches are applied to compute a set of possible illuminants. For each of them an illuminant color corrected image is evaluated on the likelihood of its semantic content: is the grass green, the road grey, and the sky blue, in correspondence with our prior knowledge of the world. The illuminant resulting in the most likely semantic composition of the image is selected as the illuminant color. To evaluate the likelihood of the semantic content, we apply probabilistic latent semantic analysis. The image is modelled as a mixture of semantic classes, such as sky, grass, road, and building. The class description is based on texture, position and color information. Experiments show that the use of high-level information improves illuminant estimation over a purely bottom-up approach. Furthermore, the proposed method is shown to significantly improve semantic class recognition performance.

Conditional Random Fields (CRFs) are an effective tool for a variety of different data segmentation and labeling tasks including visual scene interpretation, which seeks to partition images into their constituent semantic-level regions and assign appropriate class labels to each region. For accurate labeling it is important to capture the global context of the image as well as local information. We in- troduce a CRF based scene labeling model that incorporates both local features and features aggregated over the whole image or large sections of it. Secondly, traditional CRF learning requires fully labeled datasets which can be costly and troublesome to produce. We introduce a method for learning CRFs from datasets with many unlabeled nodes by marginalizing out the unknown labels so that the log-likelihood of the known ones can be maximized by gradient ascent. Loopy Belief Propagation is used to approximate the marginals needed for the gradi- ent and log-likelihood calculations and the Bethe free-energy approximation to the log-likelihood is monitored to control the step size. Our experimental results show that effective models can be learned from fragmentary labelings and that incorporating top-down aggregate features significantly improves the segmenta- tions. The resulting segmentations are compared to the state-of-the-art on three different image datasets.

There are various situations where image data is binary: character recognition, result of image segmentation etc. As a first contribution, we compare Gaussian based principal component analysis (PCA), which is often used to model images, and "binary PCA" which models the binary data more naturally using Bernoulli distributions. Furthermore, we address the problem of data alignment. Image data is often perturbed by some global transformations such as shifting, rotation, scaling etc. In such cases the data needs to be transformed to some canonical aligned form. As a second contribution, we extend the binary PCA to the "transformation invariant mixture of binary PCAs" which simultaneously corrects the data for a set of global transformations and learns the binary PCA model on the aligned data.

Appearance-based methods, based on statistical models of the pixel values in an image (region) rather than geometrical object models, are increasingly popular in computer vision. In many applications, the number of degrees of freedom (DOF) in the image generating process is much lower than the number of pixels in the image. If there is a smooth function that maps the DOF to the pixel values, then the images are confined to a low-dimensional manifold embedded in the image space. We propose a method based on probabilistic mixtures of factor analyzers to (1) model the density of images sampled from such manifolds and (2) recover global parameterizations of the manifold. A globally nonlinear probabilistic two-way mapping between coordinates on the manifold and images is obtained by combining several, locally valid, linear mappings. We propose a parameter estimation scheme that improves upon an existing scheme and experimentally compare the presented approach to self-organizing maps, generative topographic mapping, and mixtures of factor analyzers. In addition, we show that the approach also applies to finding mappings between different embeddings of the same manifold.

In this paper, we enhance appearance-based robot localization by using disparity maps. Disparity maps provide the same type of information as range based sensors (distance to objects) and thus, they are likely to be less sensitive to changes of illumination than plain images, that are the source of information generally used in appearance-based localization. The main drawback of disparity maps is that they can include very noisy depth values: points for which the algorithms can not determine reliable depth information. These noisy values have to be discarded resulting in missing values. The presence of missing values makes principal component analysis (the standard method used to compress images in the appearance-based framework) unfeasible. We describe a novel expectation-maximization algorithm to determine the principal components of a data set including missing values and we apply it to disparity maps. The results we present show that disparity maps are a valid alternative to increase the robustness of appearance-based localization.

We propose a non-linear Canonical Correlation Analysis (CCA) method which works by coordinating or aligning mixtures of linear models. In the same way that CCA extends the idea of PCA, our work extends re- cent methods for non-linear dimensionality reduction to the case where multiple embeddings of the same underlying low dimensional coordi- nates are observed, each lying on a different high dimensional manifold. We also show that a special case of our method, when applied to only a single manifold, reduces to the Laplacian Eigenmaps algorithm. As with previous alignment schemes, once the mixture models have been estimated, all of the parameters of our model can be estimated in closed form without local optima in the learning. Experimental results illustrate the viability of the approach as a non-linear extension of CCA.

This article concerns the greedy learning of gaussian mixtures. In the greedy approach, mixture components are inserted into the mixture one after the other. We propose a heuristic for searching for the optimal component to insert. In a randomized manner, a set of candidate new components is generated. For each of these candidates, we find the locally optimal new component and insert it into the existing mixture. The resulting algorithm resolves the sensitivity to initialization of state-of-the-art methods, like expectation maximization, and has running time linear in the number of data points and quadratic in the (final) number of mixture components. Due to its greedy nature, the algorithm can be particularly useful when the optimal number of mixture components is unknown. Experimental results comparing the proposed algorithm to other methods on density estimation and texture segmentation are provided.