Human Pose Estimation is a low-level task useful forsurveillance, human action recognition, and scene understandingat large. It also offers promising perspectives for the animationof synthetic characters. For all these applications, and especiallythe latter, estimating the positions of many joints is desirablefor improved performance and realism. To this purpose, wepropose a novel method called JUMPS for increasing the numberof joints in 2D pose estimates and recovering occluded ormissing joints. We believe this is the first attempt to addressthe issue. We build on a deep generative model that combines aGenerative Adversarial Network (GAN) and an encoder. TheGAN learns the distribution of high-resolution human posesequences, the encoder maps the input low-resolution sequencesto its latent space. Inpainting is obtained by computing the latentrepresentation whose decoding by the GAN generator optimallymatches the joints locations at the input. Post-processing a 2Dpose sequence using our method provides a richer representationof the character motion. We show experimentally that thelocalization accuracy of the additional joints is on average onpar with the original pose estimates.
The availability of a large labeled dataset is a key requirement for applying deep learning methods to solve various computer vision tasks. In the context of understanding human activities, existing public datasets, while large in size, are often limited to a single RGB camera and provide only per-frame or per-clip action annotations. To enable richer analysis and understanding of human activities, we introduce IKEA ASM -- a three million frame, multi-view, furniture assembly video dataset that includes depth, atomic actions, object segmentation, and human pose. Additionally, we benchmark prominent methods for video action recognition, object segmentation and human pose estimation tasks on this challenging dataset. The dataset enables the development of holistic methods, which integrate multi-modal and multi-view data to better perform on these tasks.
Human action recognition is one of the important fields of computer vision and machine learning. Although various methods have been proposed for 3D action recognition, some of which are basic and some use deep learning, the need of basic methods based on generalized eigenvalue problem is sensed for action recognition. This need is especially sensed because of having similar basic methods in the field of face recognition such as eigenfaces and Fisherfaces. In this paper, we propose Roweisposes which uses Roweis discriminant analysis for generalized subspace learning. This method includes Fisherposes, eigenposes, supervised eigenposes, and double supervised eigenposes as its special cases. Roweisposes is a family of infinite number of action recongition methods which learn a discriminative subspace for embedding the body poses. Experiments on the TST, UTKinect, and UCFKinect datasets verify the effectiveness of the proposed method for action recognition.
This technical report introduces our winning solution to the spatio-temporal action localization track, AVA-Kinetics Crossover, in ActivityNet Challenge 2020. Our entry is mainly based on Actor-Context-Actor Relation Network. We describe technical details for the new AVA-Kinetics dataset, together with some experimental results. Without any bells and whistles, we achieved 39.62 mAP on the test set of AVA-Kinetics, which outperforms other entries by a large margin. Code will be available at: https://github.com/Siyu-C/ACAR-Net.
As in many other different fields, deep learning has become the main approach in most computer vision applications, such as scene understanding, object recognition, computer-human interaction or human action recognition (HAR). Research efforts within HAR have mainly focused on how to efficiently extract and process both spatial and temporal dependencies of video sequences. In this paper, we propose and compare, two neural networks based on the convolutional long short-term memory unit, namely ConvLSTM, with differences in the architecture and the long-term learning strategy. The former uses a video-length adaptive input data generator (\emph{stateless}) whereas the latter explores the \emph{stateful} ability of general recurrent neural networks but applied in the particular case of HAR. This stateful property allows the model to accumulate discriminative patterns from previous frames without compromising computer memory. Experimental results on the large-scale NTU RGB+D dataset show that the proposed models achieve competitive recognition accuracies with lower computational cost compared with state-of-the-art methods and prove that, in the particular case of videos, the rarely-used stateful mode of recurrent neural networks significantly improves the accuracy obtained with the standard mode. The recognition accuracies obtained are 75.26\% (CS) and 75.45\% (CV) for the stateless model, with an average time consumption per video of 0.21 s, and 80.43\% (CS) and 79.91\%(CV) with 0.89 s for the stateful version.
Human actions recognition is a fundamental task in artificial vision, that has earned a great importance in recent years due to its multiple applications in different areas. %, such as the study of human behavior, security or video surveillance. In this context, this paper describes an approach for real-time human action recognition from raw depth image-sequences, provided by an RGB-D camera. The proposal is based on a 3D fully convolutional neural network, named 3DFCNN, which automatically encodes spatio-temporal patterns from depth sequences without %any costly pre-processing. Furthermore, the described 3D-CNN allows %automatic features extraction and actions classification from the spatial and temporal encoded information of depth sequences. The use of depth data ensures that action recognition is carried out protecting people's privacy% allows recognizing the actions carried out by people, protecting their privacy%\sout{of them} , since their identities can not be recognized from these data. %\st{ from depth images.} 3DFCNN has been evaluated and its results compared to those from other state-of-the-art methods within three widely used %large-scale NTU RGB+D datasets, with different characteristics (resolution, sensor type, number of views, camera location, etc.). The obtained results allows validating the proposal, concluding that it outperforms several state-of-the-art approaches based on classical computer vision techniques. Furthermore, it achieves action recognition accuracy comparable to deep learning based state-of-the-art methods with a lower computational cost, which allows its use in real-time applications.
State-of-the-art video action recognition models with complex network architecture have archived significant improvements, but these models heavily depend on large-scale well-labeled datasets. To reduce such dependency, we propose a self-supervised teacher-student architecture, i.e., the Differentiated Teachers Guided self-supervised Network (DTG-Net). In DTG-Net, except for reducing labeled data dependency by self-supervised learning (SSL), pre-trained action related models are used as teacher guidance providing prior knowledge to alleviate the demand for a large number of unlabeled videos in SSL. Specifically, leveraging the years of effort in action-related tasks, e.g., image classification, image-based action recognition, the DTG-Net learns the self-supervised video representation under various teacher guidance, i.e., those well-trained models of action-related tasks. Meanwhile, the DTG-Net is optimized in the way of contrastive self-supervised learning. When two image sequences are randomly sampled from the same video or different videos as the positive or negative pairs, respectively, they are then sent to the teacher and student networks for feature embedding. After that, the contrastive feature consistency is defined between features embedding of each pair, i.e., consistent for positive pair and inconsistent for negative pairs. Meanwhile, to reflect various teacher tasks' different guidance, we also explore different weighted guidance on teacher tasks. Finally, the DTG-Net is evaluated in two ways: (i) the self-supervised DTG-Net to pre-train the supervised action recognition models with only unlabeled videos; (ii) the supervised DTG-Net to be jointly trained with the supervised action networks in an end-to-end way. Its performance is better than most pre-training methods but also has excellent competitiveness compared to supervised action recognition methods.
In this report, we present our solution for the task of temporal action localization (detection) (task 1) in ActivityNet Challenge 2020. The purpose of this task is to temporally localize intervals where actions of interest occur and predict the action categories in a long untrimmed video. Our solution mainly includes three components: 1) feature encoding: we apply three kinds of backbones, including TSN [7], Slowfast[3] and I3d[1], which are both pretrained on Kinetics dataset[2]. Applying these models, we can extract snippet-level video representations; 2) proposal generation: we choose BMN [5] as our baseline, base on which we design a Cascade Boundary Refinement Network (CBR-Net) to conduct proposal detection. The CBR-Net mainly contains two modules: temporal feature encoding, which applies BiLSTM to encode long-term temporal information; CBR module, which targets to refine the proposal precision under different parameter settings; 3) action localization: In this stage, we combine the video-level classification results obtained by the fine tuning networks to predict the category of each proposal. Moreover, we also apply to different ensemble strategies to improve the performance of the designed solution, by which we achieve 42.788% on the testing set of ActivityNet v1.3 dataset in terms of mean Average Precision metrics.
This technical report analyzes a temporal action localization method we used in the HACS competition which is hosted in Activitynet Challenge 2020.The goal of our task is to locate the start time and end time of the action in the untrimmed video, and predict action category.Firstly, we utilize the video-level feature information to train multiple video-level action classification models. In this way, we can get the category of action in the video.Secondly, we focus on generating high quality temporal proposals.For this purpose, we apply BMN to generate a large number of proposals to obtain high recall rates. We then refine these proposals by employing a cascade structure network called Refine Network, which can predict position offset and new IOU under the supervision of ground truth.To make the proposals more accurate, we use bidirectional LSTM, Nonlocal and Transformer to capture temporal relationships between local features of each proposal and global features of the video data.Finally, by fusing the results of multiple models, our method obtains 40.55% on the validation set and 40.53% on the test set in terms of mAP, and achieves Rank 1 in this challenge.
Localizing persons and recognizing their actions from videos is a challenging task towards high-level video understanding. Recent advances have been achieved by modeling direct pairwise relations between entities. In this paper, we take one step further, not only model direct relations between pairs but also take into account indirect higher-order relations established upon multiple elements. We propose to explicitly model the Actor-Context-Actor Relation, which is the relation between two actors based on their interactions with the context. To this end, we design an Actor-Context-Actor Relation Network (ACAR-Net) which builds upon a novel High-order Relation Reasoning Operator and an Actor-Context Feature Bank to enable indirect relation reasoning for spatio-temporal action localization. Experiments on AVA and UCF101-24 datasets show the advantages of modeling actor-context-actor relations, and visualization of attention maps further verifies that our model is capable of finding relevant higher-order relations to support action detection. Notably, our method ranks first in the AVA-Kineticsaction localization task of ActivityNet Challenge 2020, out-performing other entries by a significant margin (+6.71mAP). Training code and models will be available at https://github.com/Siyu-C/ACAR-Net.
Weakly-supervised temporal action localization aims to learn detecting temporal intervals of action classes with only video-level labels. To this end, it is crucial to separate frames of action classes from the background frames (i.e., frames not belonging to any action classes). In this paper, we present a new perspective on background frames where they are modeled as out-of-distribution samples regarding their inconsistency. Then, background frames can be detected by estimating the probability of each frame being out-of-distribution, known as uncertainty, but it is infeasible to directly learn uncertainty without frame-level labels. To realize the uncertainty learning in the weakly-supervised setting, we leverage the multiple instance learning formulation. Moreover, we further introduce a background entropy loss to better discriminate background frames by encouraging their in-distribution (action) probabilities to be uniformly distributed over all action classes. Experimental results show that our uncertainty modeling is effective at alleviating the interference of background frames and brings a large performance gain without bells and whistles. We demonstrate that our model significantly outperforms state-of-the-art methods on the benchmarks, THUMOS'14 and ActivityNet (1.2 & 1.3). Our code is available at https://github.com/Pilhyeon/WTAL-Uncertainty-Modeling.
We propose a novel system for active semi-supervised feature-based action recognition. Given time sequences of features tracked during movements our system clusters the sequences into actions. Our system is based on encoder-decoder unsupervised methods shown to perform clustering by self-organization of their latent representation through the auto-regression task. These methods were tested on human action recognition benchmarks and outperformed non-feature based unsupervised methods and achieved comparable accuracy to skeleton-based supervised methods. However, such methods rely on K-Nearest Neighbours (KNN) associating sequences to actions, and general features with no annotated data would correspond to approximate clusters which could be further enhanced. Our system proposes an iterative semi-supervised method to address this challenge and to actively learn the association of clusters and actions. The method utilizes latent space embedding and clustering of the unsupervised encoder-decoder to guide the selection of sequences to be annotated in each iteration. Each iteration, the selection aims to enhance action recognition accuracy while choosing a small number of sequences for annotation. We test the approach on human skeleton-based action recognition benchmarks assuming that only annotations chosen by our method are available and on mouse movements videos recorded in lab experiments. We show that our system can boost recognition performance with only a small percentage of annotations. The system can be used as an interactive annotation tool to guide labeling efforts for 'in the wild' videos of various objects and actions to reach robust recognition.
Understanding sequential information is a fundamental task for artificial intelligence. Current neural networks attempt to learn spatial and temporal information as a whole, limited their abilities to represent large scale spatial representations over long-range sequences. Here, we introduce a new modeling strategy called Semi-Coupled Structure (SCS), which consists of deep neural networks that decouple the complex spatial and temporal concepts learning. Semi-Coupled Structure can learn to implicitly separate input information into independent parts and process these parts respectively. Experiments demonstrate that a Semi-Coupled Structure can successfully annotate the outline of an object in images sequentially and perform video action recognition. For sequence-to-sequence problems, a Semi-Coupled Structure can predict future meteorological radar echo images based on observed images. Taken together, our results demonstrate that a Semi-Coupled Structure has the capacity to improve the performance of LSTM-like models on large scale sequential tasks.
Although CNN has reached satisfactory performance in image-related tasks, using CNN to process videos is much more challenging due to the enormous size of raw video streams. In this work, we propose to use motion vectors and residuals from modern video compression techniques to effectively learn the representation of the raw frames and greatly remove the temporal redundancy, giving a faster video processing model. Compressed Video Action Recognition(CoViAR) has explored to directly use compressed video to train the deep neural network, where the motion vectors were utilized to present temporal information. However, motion vector is designed for minimizing video size where precise motion information is not obligatory. Compared with optical flow, motion vectors contain noisy and unreliable motion information. Inspired by the mechanism of video compression codecs, we propose an approach to refine the motion vectors where unreliable movement will be removed while temporal information is largely reserved. We prove that replacing the original motion vector with refined one and using the same network as CoViAR has achieved state-of-art performance on the UCF-101 and HMDB-51 with negligible efficiency degrades comparing with original CoViAR.
Industrial smoke emissions pose a significant concern to human health. Prior works have shown that using Computer Vision (CV) techniques to identify smoke as visual evidence can influence the attitude of regulators and empower citizens to pursue environmental justice. However, existing datasets are not of sufficient quality nor quantity to train the robust CV models needed to support air quality advocacy. We introduce RISE, the first large-scale video dataset for Recognizing Industrial Smoke Emissions. We adopted a citizen science approach to collaborate with local community members to annotate whether a video clip has smoke emissions. Our dataset contains 12,567 clips from 19 distinct views from cameras that monitored three industrial facilities. These daytime clips span 30 days over two years, including all four seasons. We ran experiments using deep neural networks to establish a strong performance baseline and reveal smoke recognition challenges. Our survey study discussed community feedback, and our data analysis displayed opportunities for integrating citizen scientists and crowd workers into the application of Artificial Intelligence for Social Impact.
The goal of human action recognition is to temporally or spatially localize the human action of interest in video sequences. Temporal localization (i.e. indicating the start and end frames of the action in a video) is referred to as frame-level detection. Spatial localization, which is more challenging, means to identify the pixels within each action frame that correspond to the action. This setting is usually referred to as pixel-level detection. In this chapter, we are using action, activity, event interchangeably.
The popular task of 3D human action recognition is almost exclusively solved by training deep-learning classifiers. To achieve a high recognition accuracy, the input 3D actions are often pre-processed by various normalization or augmentation techniques. However, it is not computationally feasible to train a classifier for each possible variant of training data in order to select the best-performing subset of pre-processing techniques for a given dataset. In this paper, we propose to train an independent classifier for each available pre-processing technique and fuse the classification results based on a strict majority vote rule. Together with a proposed evaluation procedure, we can very efficiently determine the best combination of normalization and augmentation techniques for a specific dataset. For the best-performing combination, we can retrospectively apply the normalized/augmented variants of input data to train only a single classifier. This also allows us to decide whether it is better to train a single model, or rather a set of independent classifiers.
Residual networks (ResNets) represent a powerful type of convolutional neural network (CNN) architecture, widely adopted and used in various tasks. In this work we propose an improved version of ResNets. Our proposed improvements address all three main components of a ResNet: the flow of information through the network layers, the residual building block, and the projection shortcut. We are able to show consistent improvements in accuracy and learning convergence over the baseline. For instance, on ImageNet dataset, using the ResNet with 50 layers, for top-1 accuracy we can report a 1.19% improvement over the baseline in one setting and around 2% boost in another. Importantly, these improvements are obtained without increasing the model complexity. Our proposed approach allows us to train extremely deep networks, while the baseline shows severe optimization issues. We report results on three tasks over six datasets: image classification (ImageNet, CIFAR-10 and CIFAR-100), object detection (COCO) and video action recognition (Kinetics-400 and Something-Something-v2). In the deep learning era, we establish a new milestone for the depth of a CNN. We successfully train a 404-layer deep CNN on the ImageNet dataset and a 3002-layer network on CIFAR-10 and CIFAR-100, while the baseline is not able to converge at such extreme depths. Code is available at: https://github.com/iduta/iresnet
Despite the success in still image recognition, deep neural networks for spatiotemporal signal tasks (such as human action recognition in videos) still suffers from low efficacy and inefficiency over the past years. Recently, human experts have put more efforts into analyzing the importance of different components in 3D convolutional neural networks (3D CNNs) to design more powerful spatiotemporal learning backbones. Among many others, spatiotemporal fusion is one of the essentials. It controls how spatial and temporal signals are extracted at each layer during inference. Previous attempts usually start by ad-hoc designs that empirically combine certain convolutions and then draw conclusions based on the performance obtained by training the corresponding networks. These methods only support network-level analysis on limited number of fusion strategies. In this paper, we propose to convert the spatiotemporal fusion strategies into a probability space, which allows us to perform network-level evaluations of various fusion strategies without having to train them separately. Besides, we can also obtain fine-grained numerical information such as layer-level preference on spatiotemporal fusion within the probability space. Our approach greatly boosts the efficiency of analyzing spatiotemporal fusion. Based on the probability space, we further generate new fusion strategies which achieve the state-of-the-art performance on four well-known action recognition datasets.
Skeleton data, which consists of only the 2D/3D coordinates of the human joints, has been widely studied for human action recognition. Existing methods take the semantics as prior knowledge to group human joints and draw correlations according to their spatial locations, which we call the semantic perspective for skeleton modeling. In this paper, in contrast to previous approaches, we propose to model skeletons from a novel spatial perspective, from which the model takes the spatial location as prior knowledge to group human joints and mines the discriminative patterns of local areas in a hierarchical manner. The two perspectives are orthogonal and complementary to each other; and by fusing them in a unified framework, our method achieves a more comprehensive understanding of the skeleton data. Besides, we customized two networks for the two perspectives. From the semantic perspective, we propose a Transformer-like network that is expert in modeling joint correlations, and present three effective techniques to adapt it for skeleton data. From the spatial perspective, we transform the skeleton data into the sparse format for efficient feature extraction and present two types of sparse convolutional networks for sparse skeleton modeling. Extensive experiments are conducted on three challenging datasets for skeleton-based human action/gesture recognition, namely, NTU-60, NTU-120 and SHREC, where our method achieves state-of-the-art performance.