This article has been retracted: please see Elsevier policy on article withdrawal ( https://www.elsevier.com/about/policies-and-standards/article-withdrawal ). This article has been retracted at the request of the authors. The authors reported critical technical concerns caused by the ignored parameter Θ (a dynamic threshold that controls the specific privacy requirements of the used dataset) in the experiment evaluation process. Therefore, the authors requested retraction, and the Editors agreed with this request. The authors regret the error.

Inspired by the human ability to infer emotions from body language, we propose an automated framework for body language based emotion recognition starting from regular RGB videos. In collaboration with psychologists, we further extend the framework for psychiatric symptom prediction. Because a specific application domain of the proposed framework may only supply a limited amount of data, the framework is designed to work on a small training set and possess a good transferability. The proposed system in the first stage generates sequences of body language predictions based on human poses estimated from input videos. In the second stage, the predicted sequences are fed into a temporal network for emotion interpretation and psychiatric symptom prediction. We first validate the accuracy and transferability of the proposed body language recognition method on several public action recognition datasets. We then evaluate the framework on a proposed URMC dataset, which consists of conversations between a standardized patient and a behavioral health professional, along with expert annotations of body language, emotions, and potential psychiatric symptoms. The proposed framework outperforms other methods on the URMC dataset.

Human body part segmentation refers to the task of predicting the semantic segmentation mask for each body part. Fully supervised body part segmentation methods achieve good performances but require an enormous amount of effort to annotate part masks for training. In contrast to high annotation costs needed for a limited number of part mask annotations, a large number of weak labels such as poses and full body masks already exist and contain relevant information. Motivated by the possibility of using existing weak labels, we propose the first weakly supervised body part segmentation framework. The core idea is first converting the sparse weak labels such as keypoints to the initial estimate of body part masks, and then iteratively refine the part mask predictions. We name the initial part masks estimated from poses the “part priors”. with sufficient extra weak labels, our weakly supervised framework achieves a comparable performance (62.0% mIoU) to the fully supervised method (63.6% mIoU) on the Pascal-Person-Part dataset. Furthermore, in the extended semi-supervised setting, the proposed framework outperforms the state-of-art methods. Moreover, we extend our proposed framework to other keypoint-supervised part segmentation tasks such as face parsing.

This work addresses a novel and challenging problem of estimating the full 3D hand shape and pose from a single RGB image. Most current methods in 3D hand analysis from monocular RGB images only focus on estimating the 3D locations of hand keypoints, which cannot fully express the 3D shape of hand. In contrast, we propose a Graph Convolutional Neural Network (Graph CNN) based method to reconstruct a full 3D mesh of hand surface that contains richer information of both 3D hand shape and pose. To train networks with full supervision, we create a large-scale synthetic dataset containing both ground truth 3D meshes and 3D poses. When fine-tuning the networks on real-world datasets without 3D ground truth, we propose a weakly-supervised approach by leveraging the depth map as a weak supervision in training. Through extensive evaluations on our proposed new datasets and two public datasets, we show that our proposed method can produce accurate and reasonable 3D hand mesh, and can achieve superior 3D hand pose estimation accuracy when compared with state-of-the-art methods.

Existing video-based human pose estimation methods extensively apply large networks onto every frame in the video to localize body joints, which suffer high computational cost and hardly meet the low-latency requirement in realistic applications. To address this issue, we propose a novel Dynamic Kernel Distillation (DKD) model to facilitate small networks for estimating human poses in videos, thus significantly lifting the efficiency. In particular, DKD introduces a light-weight distillator to online distill pose kernels via leveraging temporal cues from the previous frame in a one-shot feed-forward manner. Then, DKD simplifies body joint localization into a matching procedure between the pose kernels and the current frame, which can be efficiently computed via simple convolution. In this way, DKD fast transfers pose knowledge from one frame to provide compact guidance for body joint localization in the following frame, which enables utilization of small networks in video-based pose estimation. To facilitate the training process, DKD exploits a temporally adversarial training strategy that introduces a temporal discriminator to help generate temporally coherent pose kernels and pose estimation results within a long range. Experiments on Penn Action and Sub-JHMDB benchmarks demonstrate outperforming efficiency of DKD, specifically, 10x flops reduction and 2x speedup over previous best model, and its state-of-the-art accuracy.

Landmark localization in images and videos is a classic problem solved in various ways. Nowadays, with deep networks prevailing throughout machine learning, there are revamped interests in pushing facial landmark detectors to handle more challenging data. Most efforts use network objectives based on L1 or L2 norms, which have several disadvantages. First of all, the generated heatmaps translate to the locations of landmarks (i.e. confidence maps) from which predicted landmark locations (i.e. the means) get penalized without accounting for the spread: a high- scatter corresponds to low confidence and vice-versa. For this, we introduce a LaplaceKL objective that penalizes for low confidence. Another issue is a dependency on labeled data, which are expensive to obtain and susceptible to error. To address both issues, we propose an adversarial training framework that leverages unlabeled data to improve model performance. Our method claims state-of-the-art on all of the 300W benchmarks and ranks second-to-best on the Annotated Facial Landmarks in the Wild (AFLW) dataset. Furthermore, our model is robust with a reduced size: 1/8 the number of channels (i.e. 0.0398 MB) is comparable to the state-of-the-art in real-time on CPU. Thus, this work is of high practical value to real-life application.

Different from the fully-supervised action detection problem that is dependent on expensive frame-level annotations, weakly supervised action detection (WSAD) only needs video-level annotations, making it more practical for real-world applications. Existing WSAD methods detect action instances by scoring each video segment (a stack of frames) individually. Most of them fail to model the temporal relations among video segments and cannot effectively characterize action instances possessing latent temporal structure. To alleviate this problem in WSAD, we propose the temporal structure mining (TSM) approach. In TSM, each action instance is modeled as a multi-phase process and phase evolving within an action instance, \emph{i.e.}, the temporal structure, is exploited. Meanwhile, the video background is modeled by a background phase, which separates different action instances in an untrimmed video. In this framework, phase filters are used to calculate the confidence scores of the presence of an action's phases in each segment. Since in the WSAD task, frame-level annotations are not available and thus phase filters cannot be trained directly. To tackle the challenge, we treat each segment's phase as a hidden variable. We use segments' confidence scores from each phase filter to construct a table and determine hidden variables, i.e., phases of segments, by a maximal circulant path discovery along the table. Experiments conducted on three benchmark datasets demonstrate the state-of-the-art performance of the proposed TSM.

Action recognition with 3D skeleton sequences became popular due to its speed and robustness. The recently proposed convolutional neural networks (CNNs)-based methods show a good performance in learning spatio-temporal representations for skeleton sequences. Despite the good recognition accuracy achieved by previous CNN-based methods, there existed two problems that potentially limit the performance. First, previous skeleton representations were generated by chaining joints with a fixed order. The corresponding semantic meaning was unclear and the structural information among the joints was lost. Second, previous models did not have an ability to focus on informative joints. The attention mechanism was important for skeleton-based action recognition because different joints contributed unequally toward the correct recognition. To solve these two problems, we proposed a novel CNN-based method for skeleton-based action recognition. We first redesigned the skeleton representations with a depth-first tree traversal order, which enhanced the semantic meaning of skeleton images and better preserved the associated structural information. We then proposed the general two-branch attention architecture that automatically focused on spatio-temporal key stages and filtered out unreliable joint predictions. Based on the proposed general architecture, we designed a global long-sequence attention network with refined branch structures. Furthermore, in order to adjust the kernel's spatio-temporal aspect ratios and better capture long-term dependencies, we proposed a sub-sequence attention network (SSAN) that took sub-image sequences as inputs. We showed that the two-branch attention architecture could be combined with the SSAN to further improve the performance. Our experiment results on the NTU RGB+D data set and the SBU kinetic interaction data set outperformed the state of the art. The model was further validated on noisy estimated poses from the subsets of the UCF101 data set and the kinetics data set.

Action recognition with 3D skeleton sequences is becoming popular due to its speed and robustness. The recently proposed Convolutional Neural Networks (CNN) based methods have shown good performance in learning spatio-temporal representations for skeleton sequences. Despite the good recognition accuracy achieved by previous CNN based methods, there exist two problems that potentially limit the performance. First, previous skeleton representations are generated by chaining joints with a fixed order. The corresponding semantic meaning is unclear and the structural information among the joints is lost. Second, previous models do not have an ability to focus on informative joints. The attention mechanism is important for skeleton based action recognition because there exist spatio-temporal key stages and the joint predictions can be inaccurate. To solve the two problems, we propose a novel CNN based method for skeleton based action recognition. We first redesign the skeleton representations with a depth-first tree traversal order, which enhances the semantic meaning of skeleton images and better preserves the structural information. We then propose the idea of a two-branch attention architecture that focuses on spatio-temporal key stages and filters out unreliable joint predictions. A base attention model with the simplest structure is first introduced to illustrate the two-branch attention architecture. By improving the structures in both branches, we further propose a Global Long-sequence Attention Network (GLAN). Experiment results on the NTU RGB+D dataset and the SBU Kinetic Interaction dataset show that our proposed approach outperforms the state-of-the-art, as well as the effectiveness of each component.

Learning to rank has recently emerged as an attractive technique to train deep convolutional neural networks for various computer vision tasks. Pairwise ranking, in particular, has been successful in multi-label image classification, achieving state-of-the-art results on various benchmarks. However, most existing approaches use the hinge loss to train their models, which is non-smooth and thus is difficult to optimize especially with deep networks. Furthermore, they employ simple heuristics, such as top-k or thresholding, to determine which labels to include in the output from a ranked list of labels, which limits their use in the real-world setting. In this work, we propose two techniques to improve pairwise ranking based multi-label image classification by solving the aforementioned problems: (1) we propose a novel loss function for pairwise ranking, which is smooth everywhere, and (2) we incorporate a label decision module into the model, estimating the optimal confidence thresholds for each visual concept. We provide theoretical analyses of our loss function from the point of view of the Bayes consistency and risk minimization, and show its benefit over existing pairwise ranking formulations. We also demonstrate the effectiveness of our approach on two large-scale datasets, NUS-WIDE and MS-COCO, achieving the best reported result in the literature.

The ability of learning from noisy labels is very useful in many visual recognition tasks, as a vast amount of data with noisy labels are relatively easy to obtain. Traditionally, label noise has been treated as statistical outliers, and techniques such as importance re-weighting and bootstrapping have been proposed to alleviate the problem. According to our observation, the real-world noisy labels exhibit multimode characteristics as the true labels, rather than behaving like independent random outliers. In this work, we propose a unified distillation framework to use “side” information, including a small clean dataset and label relations in knowledge graph, to “hedge the risk” of learning from noisy labels. Unlike the traditional approaches evaluated based on simulated label noises, we propose a suite of new benchmark datasets, in Sports, Species and Artifacts domains, to evaluate the task of learning from noisy labels in the practical setting. The empirical study demonstrates the effectiveness of our proposed method in all the domains.

Initializing an effective dictionary is an indispensable step for sparse representation. In this paper, we focus on the dictionary selection problem with the objective to select a compact subset of basis from original training data instead of learning a new dictionary matrix as dictionary learning models do. We first design a new dictionary selection model via l2,0norm. For model optimization, we propose two methods: one is the standard forward-backward greedy algorithm, which is not suitable for large-scale problems; the other is based on the gradient cues at each forward iteration and speeds up the process dramatically. In comparison with the state-of-the-art dictionary selection models, our model is not only more effective and efficient, but also can control the sparsity. To evaluate the performance of our new model, we select two practical web media summarization problems: 1) we build a new data set consisting of around 500 users, 3000 albums, and 1 million images, and achieve effective assisted albuming based on our model and 2) by formulating the video summarization problem as a dictionary selection issue, we employ our model to extract keyframes from a video sequence in a more flexible way. Generally, our model outperforms the state-of-the-art methods in both these two tasks.

Composing fashion outfits involves deep under-standing of fashion standards while incorporating creativity for choosing multiple fashion items (e.g., jewelry, bag, pants, dress). In fashion websites, popular or high-quality fashion outfits are usually designed by fashion experts and followed by large audiences. In this paper, we propose a machine learning system to compose fashion outfits automatically. The core of the proposed automatic composition system is to score fashion outfit candidates based on the appearances and metadata. We propose to leverage outfit popularity on fashion-oriented websites to supervise the scoring component. The scoring component is a multimodal multiinstance deep learning system that evaluates instance aesthetics and set compatibility simultaneously. In order to train and evaluate the proposed composition system, we have collected a large-scale fashion outfit dataset with 195K outfits and 368K fashion items from Polyvore. Although the fashion outfit scoring and composition is rather challenging, we have achieved an AUC of 85% for the scoring component, and an accuracy of 77% for a constrained composition task.

With the recent popularity of animated GIFs on social media, there is need for ways to index them with rich meta-data. To advance research on animated GIF understanding, we collected a new dataset, Tumblr GIF (TGIF), with 100K animated GIFs from Tumblr and 120K natural language descriptions obtained via crowdsourcing. The motivation for this work is to develop a testbed for image sequence description systems, where the task is to generate natural language descriptions for animated GIFs or video clips. To ensure a high quality dataset, we developed a series of novel quality controls to validate free-form text input from crowd-workers. We show that there is unambiguous association between visual content and natural language descriptions in our dataset, making it an ideal benchmark for the visual content captioning task. We perform extensive statistical analyses to compare our dataset to existing image and video description datasets. Next, we provide baseline results on the animated GIF description task, using three representative techniques: nearest neighbor, statistical machine translation, and recurrent neural networks. Finally, we show that models fine-tuned from our animated GIF description dataset can be helpful for automatic movie description.

Human gait is an important biometric feature for person identification in surveillance videos because it can be collected at a distance without subject cooperation. Most existing gait recognition methods are based on Gait Energy Image (GEI). Although the spatial information in one gait sequence can be well represented by GEI, the temporal information is lost. To solve this problem, we propose a new feature learning method for gait recognition. Not only can the learned feature preserve temporal information in a gait sequence, but it can also be applied to cross-view gait recognition. Heatmaps extracted by a convolutional neutral network (CNN) based pose estimate method are used to describe the gait information in one frame. To model a gait sequence, the LSTM recurrent neural network is naturally adopted. Our LSTM model can be trained with unlabeled data, where the identity of the subject in a gait sequence is unknown. When labeled data are available, our LSTM works as a frame to frame view transformation model (VTM). Experiments on a gait benchmark demonstrate the efficacy of our method.

In this study, we investigate what a practically useful approach is in order to achieve robust skin disease diagnosis. A direct approach is to target the ground truth diagnosis labels, while an alternative approach instead focuses on determining skin lesion characteristics that are more visually consistent and discernible. We argue that, for computer aided skin disease diagnosis, it is both more realistic and more useful that lesion type tags should be considered as the target of an automated diagnosis system such that the system can first achieve a high accuracy in describing skin lesions, and in turn facilitate disease diagnosis using lesion characteristics in conjunction with other evidences. To further meet such an objective, we employ convolutional neutral networks (CNNs) for both the disease-targeted and lesion-targeted classifications. We have collected a large-scale and diverse dataset of 75,665 skin disease images from six publicly available dermatology atlantes. Then we train and compare both disease-targeted and lesion-targeted classifiers, respectively. For disease-targeted classification, only 27.6% top-1 accuracy and 57.9% top-5 accuracy are achieved with a mean average precision (mAP) of 0.42. In contrast, for lesion-targeted classification, we can achieve a much higher mAP of 0.70.

This paper considers the problem of fine-grained image recognition with a growing vocabulary. Since in many real world applications we often have to add a new object category or visual concept with just a few images to learn from, it is crucial to develop a method that is able to generalize the recognition model from existing classes to new classes. Deep convolutional neural networks are capable of constructing powerful image representations; however, these networks usually rely on a logistic loss function that cannot handle the incremental learning problem. In this paper, we present a new method that can efficiently learn a new class given only a limited number of training examples, which we evaluate on the problems of food and clothing recognition. To illustrate the performance of our proposed method on the task of recognizing different kinds of food, when using only 1.3\% of training examples per category we achieved about 73\% of the performance (as measured by F1-score) compared to when using all available training data.

Most state-of-the-art image retrieval and recommendation systems predominantly focus on individual images. In contrast, socially curated image collections, condensing distinctive yet coherent images into one set, are largely overlooked by the research communities. In this paper, we aim to design a novel recommendation system that can provide users with image collections relevant to individual personal preferences and interests. To this end, two key issues need to be addressed, i.e., image collection modeling and similarity measurement. For image collection modeling, we consider each image collection as a whole in a group sparse reconstruction framework and extract concise collection descriptors given the pretrained dictionaries. We then consider image collection recommendation as a dynamic similarity measurement problem in response to user's clicked image set, and employ a metric learner to measure the similarity between the image collection and the clicked image set. As there is no previous work directly comparable to this study, we implement several competitive baselines and related methods for comparison. The evaluations on a large scale Pinterest data set have validated the effectiveness of our proposed methods for modeling and recommending image collections.

Video entity linking, which connects online videos to the related entities in a semantic knowledge base, can enable a wide variety of video based applications including video retrieval and video recommendation. Most existing systems for video entity linking rely on video metadata. In this paper, we propose to exploit video visual content to improve video entity linking. In the proposed framework, videos are first linked to entity candidates using a text-based method. Next, the entity candidates are verified and reranked according to visual content. In order to properly handle large variations in visual content matching, we propose to use Multiple Instance Metric Learning to learn a "set to sequence" metric for this specific matching problem. To evaluate the proposed framework, we collect and annotate 1912 videos crawled from the YouTube open API. Experiment results have shown consistent gains by the proposed framework over several strong baselines.

Pinboard on Pinterest is an emerging media to engage online social media users, on which users post online images for specific topics. Regardless of its significance, there is little previous work specifically to facilitate information discovery based on pinboards. This paper proposes a novel pinboard recommendation system for Twitter users. In order to associate contents from the two social media platforms, we propose to use MultiLabel classification to map Twitter user followees to pinboard topics and visual diversification to recommend pinboards given user interested topics. A preliminary experiment on a dataset with 2000 users validated our proposed system.

The asynchronous parallel implementations of stochastic gradient (SG) have been broadly used in solving deep neural network and received many successes in practice recently. However, existing theories cannot explain their convergence and speedup properties, mainly due to the nonconvexity of most deep learning formulations and the asynchronous parallel mechanism. To fill the gaps in theory and provide theoretical supports, this paper studies two asynchronous parallel implementations of SG: one is on the computer network and the other is on the shared memory system. We establish an ergodic convergence rate $O(1/\sqrt{K})$ for both algorithms and prove that the linear speedup is achievable if the number of workers is bounded by $\sqrt{K}$ ($K$ is the total number of iterations). Our results generalize and improve existing analysis for convex minimization.

We introduce and investigate a novel problem of image recommendation for web search engine users. Modern web search engines have become a critical assistant for people's daily life. Through interacting with web search engines, users exhibit personalized information needs in various aspects. While this information is critical to improve user experience, it is mostly used only in the web search domain. In this paper, we propose to leverage web search engine users' behavior data to perform image recommendation. To this end, we have developed a two-stage method to label users' preferences for images through crowdsourcing techniques. The two-stage annotation consists of 1) inferring a user's general interests and 2) estimating if this user will be interested in an image. In addition, we implement a baseline algorithm to demonstrate the promise of the proposed cross-domain recommendation framework.

Within the context of the Learning from Narrated Demonstration framework, a key vision component is to detect the task-relevant object for further processing. In this paper, we take advantage of the fact the task-relevant object is often connected to the supervisor's hand and recast the problem as handheld object detection and tracking. Achieving robust handheld object detection and tracking has its own challenges, including arbitrary object appearance, viewpoint and non-rigid deformation. We propose a robust vision system that integrates speech information to perform handheld object detection via CRF, and MeanShift based tracking. Extensive evaluation on five sets of data has demonstrated the validity and robustness of the proposed system.

We present a general framework for complex event recognition that is well-suited for integrating information that varies widely in detail and granularity. Consider the scenario of an agent in an instrumented space performing a complex task while describing what he is doing in a natural manner. The system takes in a variety of information, including objects and gestures recognized by RGB-D and descriptions of events extracted from recognized and parsed speech. The system outputs a complete reconstruction of the agent's plan, explaining actions in terms of more complex activities and filling in unobserved but necessary events. We show how to use Markov Logic (a probabilistic extension of first-order logic) to create a model in which observations can be partial, noisy, and refer to future or temporally ambiguous events; complex events are composed from simpler events in a manner that exposes their structure for inference and learning; and uncertainty is handled in a sound probabilistic manner. We demonstrate the effectiveness of the approach for tracking kitchen activities in the presence of noisy and incomplete observations.