In recent years, the performance of action recognition has been significantly improved with the help of deep neural networks. Most of the existing action recognition works hold the \textit{closed-set} assumption that all action categories are known beforehand while deep networks can be well trained for these categories. However, action recognition in the real world is essentially an \textit{open-set} problem, namely, it is impossible to know all action categories beforehand and consequently infeasible to prepare sufficient training samples for those emerging categories. In this case, applying closed-set recognition methods will definitely lead to unseen-category errors. To address this challenge, we propose the Open Deep Network (ODN) for the open-set action recognition task. Technologically, ODN detects new categories by applying a multi-class triplet thresholding method, and then dynamically reconstructs the classification layer and "opens" the deep network by adding predictors for new categories continually. In order to transfer the learned knowledge to the new category, two novel methods, Emphasis Initialization and Allometry Training, are adopted to initialize and incrementally train the new predictor so that only few samples are needed to fine-tune the model. Extensive experiments show that ODN can effectively detect and recognize new categories with little human intervention, thus applicable to the open-set action recognition tasks in the real world. Moreover, ODN can even achieve comparable performance to some closed-set methods.
In visual surveillance systems, it is necessary to recognize the behavior of people handling objects such as a phone, a cup, or a plastic bag. In this paper, to address this problem, we propose a new framework for recognizing object-related human actions by graph convolutional networks using human and object poses. In this framework, we construct skeletal graphs of reliable human poses by selectively sampling the informative frames in a video, which include human joints with high confidence scores obtained in pose estimation. The skeletal graphs generated from the sampled frames represent human poses related to the object position in both the spatial and temporal domains, and these graphs are used as inputs to the graph convolutional networks. Through experiments over an open benchmark and our own data sets, we verify the validity of our framework in that our method outperforms the state-of-the-art method for skeleton-based action recognition.
In this paper, we propose the use of a semantic image, an improved representation for video analysis, principally in combination with Inception networks. The semantic image is obtained by applying localized sparse segmentation using global clustering (LSSGC) prior to the approximate rank pooling which summarizes the motion characteristics in single or multiple images. It incorporates the background information by overlaying a static background from the window onto the subsequent segmented frames. The idea is to improve the action-motion dynamics by focusing on the region which is important for action recognition and encoding the temporal variances using the frame ranking method. We also propose the sequential combination of Inception-ResNetv2 and long-short-term memory network (LSTM) to leverage the temporal variances for improved recognition performance. Extensive analysis has been carried out on UCF101 and HMDB51 datasets which are widely used in action recognition studies. We show that (i) the semantic image generates better activations and converges faster than its original variant, (ii) using segmentation prior to approximate rank pooling yields better recognition performance, (iii) The use of LSTM leverages the temporal variance information from approximate rank pooling to model the action behavior better than the base network, (iv) the proposed representations can be adaptive as they can be used with existing methods such as temporal segment networks to improve the recognition performance, and (v) our proposed four-stream network architecture comprising of semantic images and semantic optical flows achieves state-of-the-art performance, 95.9% and 73.5% recognition accuracy on UCF101 and HMDB51, respectively.
Motion has shown to be useful for video understanding, where motion is typically represented by optical flow. However, computing flow from video frames is very time-consuming. Recent works directly leverage the motion vectors and residuals readily available in the compressed video to represent motion at no cost. While this avoids flow computation, it also hurts accuracy since the motion vector is noisy and has substantially reduced resolution, which makes it a less discriminative motion representation. To remedy these issues, we propose a lightweight generator network, which reduces noises in motion vectors and captures fine motion details, achieving a more Discriminative Motion Cue (DMC) representation. Since optical flow is a more accurate motion representation, we train the DMC generator to approximate flow using a reconstruction loss and a generative adversarial loss, jointly with the downstream action classification task. Extensive evaluations on three action recognition benchmarks (HMDB-51, UCF-101, and a subset of Kinetics) confirm the effectiveness of our method. Our full system, consisting of the generator and the classifier, is coined as DMC-Net which obtains high accuracy close to that of using flow and runs two orders of magnitude faster than using optical flow at inference time.
In this paper, we deal with the problem of temporal action localization for a large-scale untrimmed cricket videos dataset. Our action of interest for cricket videos is a cricket stroke played by a batsman, which is, usually, covered by cameras placed at the stands of the cricket ground at both ends of the cricket pitch. After applying a sequence of preprocessing steps, we have ~73 million frames for 1110 videos in the dataset at constant frame rate and resolution. The method of localization is a generalized one which applies a trained random forest model for CUTs detection(using summed up grayscale histogram difference features) and two linear SVM camera models(CAM1 and CAM2) for first frame detection, trained on HOG features of CAM1 and CAM2 video shots. CAM1 and CAM2 are assumed to be part of the cricket stroke. At the predicted boundary positions, the HOG features of the first frames are computed and a simple algorithm was used to combine the positively predicted camera shots. In order to make the process as generic as possible, we did not consider any domain specific knowledge, such as tracking or specific shape and motion features. The detailed analysis of our methodology is provided along with the metrics used for evaluation of individual models, and the final predicted segments. We achieved a weighted mean TIoU of 0.5097 over a small sample of the test set.
Current UAV-recorded datasets are mostly limited to action recognition and object tracking, whereas the gesture signals datasets were mostly recorded in indoor spaces. Currently, there is no outdoor recorded public video dataset for UAV commanding signals. Gesture signals can be effectively used with UAVs by leveraging the UAVs visual sensors and operational simplicity. To fill this gap and enable research in wider application areas, we present a UAV gesture signals dataset recorded in an outdoor setting. We selected 13 gestures suitable for basic UAV navigation and command from general aircraft handling and helicopter handling signals. We provide 119 high-definition video clips consisting of 37151 frames. The overall baseline gesture recognition performance computed using Pose-based Convolutional Neural Network (P-CNN) is 91.9 %. All the frames are annotated with body joints and gesture classes in order to extend the dataset's applicability to a wider research area including gesture recognition, action recognition, human pose recognition and situation awareness.
In this work, we address two coupled tasks of gaze prediction and action recognition in egocentric videos by exploring their mutual context. Our assumption is that in the procedure of performing a manipulation task, what a person is doing determines where the person is looking at, and the gaze point reveals gaze and non-gaze regions which contain important and complementary information about the undergoing action. We propose a novel mutual context network (MCN) that jointly learns action-dependent gaze prediction and gaze-guided action recognition in an end-to-end manner. Experiments on public egocentric video datasets demonstrate that our MCN achieves state-of-the-art performance of both gaze prediction and action recognition.
We describe a novel cross-modal embedding space for actions, named Action2Vec, which combines linguistic cues from class labels with spatio-temporal features derived from video clips. Our approach uses a hierarchical recurrent network to capture the temporal structure of video features. We train our embedding using a joint loss that combines classification accuracy with similarity to Word2Vec semantics. We evaluate Action2Vec by performing zero shot action recognition and obtain state of the art results on three standard datasets. In addition, we present two novel analogy tests which quantify the extent to which our joint embedding captures distributional semantics. This is the first joint embedding space to combine verbs and action videos, and the first to be thoroughly evaluated with respect to its distributional semantics.
Advanced video classification systems decode video frames to derive the necessary texture and motion representations for ingestion and analysis by spatio-temporal deep convolutional neural networks (CNNs). However, when considering visual Internet-of-Things applications, surveillance systems and semantic crawlers of large video repositories, the video capture and the CNN-based semantic analysis parts do not tend to be co-located. This necessitates the transport of compressed video over networks and incurs significant overhead in bandwidth and energy consumption, thereby significantly undermining the deployment potential of such systems. In this paper, we investigate the trade-off between the encoding bitrate and the achievable accuracy of CNN-based video classification models that directly ingest AVC/H.264 and HEVC encoded videos. Instead of retaining entire compressed video bitstreams and applying complex optical flow calculations prior to CNN processing, we only retain motion vector and select texture information at significantly-reduced bitrates and apply no additional processing prior to CNN ingestion. Based on three CNN architectures and two action recognition datasets, we achieve 11%-94% saving in bitrate with marginal effect on classification accuracy. A model-based selection between multiple CNNs increases these savings further, to the point where, if up to 7% loss of accuracy can be tolerated, video classification can take place with as little as 3 kbps for the transport of the required compressed video information to the system implementing the CNN models.
Recognizing human actions in untrimmed videos is an important challenging task. An effective 3D motion representation and a powerful learning model are two key factors influencing recognition performance. In this paper we introduce a new skeleton-based representation for 3D action recognition in videos. The key idea of the proposed representation is to transform 3D joint coordinates of the human body carried in skeleton sequences into RGB images via a color encoding process. By normalizing the 3D joint coordinates and dividing each skeleton frame into five parts, where the joints are concatenated according to the order of their physical connections, the color-coded representation is able to represent spatio-temporal evolutions of complex 3D motions, independently of the length of each sequence. We then design and train different Deep Convolutional Neural Networks (D-CNNs) based on the Residual Network architecture (ResNet) on the obtained image-based representations to learn 3D motion features and classify them into classes. Our method is evaluated on two widely used action recognition benchmarks: MSR Action3D and NTU-RGB+D, a very large-scale dataset for 3D human action recognition. The experimental results demonstrate that the proposed method outperforms previous state-of-the-art approaches whilst requiring less computation for training and prediction.
Recognizing actions in ice hockey using computer vision poses challenges due to bulky equipment and inadequate image quality. A novel two-stream framework has been designed to improve action recognition accuracy for hockey using three main components. First, pose is estimated via the Part Affinity Fields model to extract meaningful cues from the player. Second, optical flow (using LiteFlowNet) is used to extract temporal features. Third, pose and optical flow streams are fused and passed to fully-connected layers to estimate the hockey player's action. A novel publicly available dataset named HARPET (Hockey Action Recognition Pose Estimation, Temporal) was created, composed of sequences of annotated actions and pose of hockey players including their hockey sticks as an extension of human body pose. Three contributions are recognized. (1) The novel two-stream architecture achieves 85% action recognition accuracy, with the inclusion of optical flows increasing accuracy by about 10%. (2) The unique localization of hand-held objects (e.g., hockey sticks) as part of pose increases accuracy by about 13%. (3) For pose estimation, a bigger and more general dataset, MSCOCO, is successfully used for transfer learning to a smaller and more specific dataset, HARPET, achieving a PCKh of 87%.
Many semantic video analysis tasks can benefit from multiple, heterogenous signals. For example, in addition to the original RGB input sequences, sequences of optical flow are usually used to boost the performance of human action recognition in videos. To learn from these heterogenous input sources, existing methods reply on two-stream architectural designs that contain independent, parallel streams of Recurrent Neural Networks (RNNs). However, two-stream RNNs do not fully exploit the reciprocal information contained in the multiple signals, let alone exploit it in a recurrent manner. To this end, we propose in this paper a novel recurrent architecture, termed Coupled Recurrent Network (CRN), to deal with multiple input sources. In CRN, the parallel streams of RNNs are coupled together. Key design of CRN is a Recurrent Interpretation Block (RIB) that supports learning of reciprocal feature representations from multiple signals in a recurrent manner. Different from RNNs which stack the training loss at each time step or the last time step, we propose an effective and efficient training strategy for CRN. Experiments show the efficacy of the proposed CRN. In particular, we achieve the new state of the art on the benchmark datasets of human action recognition and multi-person pose estimation.
Canonical correlation analysis (CCA) is a statistical learning method that seeks to build view-independent latent representations from multi-view data. This method has been successfully applied to several pattern analysis tasks such as image-to-text mapping and view-invariant object/action recognition. However, this success is highly dependent on the quality of data pairing (i.e., alignments) and mispairing adversely affects the generalization ability of the learned CCA representations. In this paper, we address the issue of alignment errors using a new variant of canonical correlation analysis referred to as alignment-agnostic (AA) CCA. Starting from erroneously paired data taken from different views, this CCA finds transformation matrices by optimizing a constrained maximization problem that mixes a data correlation term with context regularization; the particular design of these two terms mitigates the effect of alignment errors when learning the CCA transformations. Experiments conducted on multi-view tasks, including multi-temporal satellite image change detection, show that our AA CCA method is highly effective and resilient to mispairing errors.
State-of-the-art methods for video action recognition commonly use an ensemble of two networks: the spatial stream, which takes RGB frames as input, and the temporal stream, which takes optical flow as input. In recent work, both of these streams consist of 3D Convolutional Neural Networks, which apply spatiotemporal filters to the video clip before performing classification. Conceptually, the temporal filters should allow the spatial stream to learn motion representations, making the temporal stream redundant. However, we still see significant benefits in action recognition performance by including an entirely separate temporal stream, indicating that the spatial stream is "missing" some of the signal captured by the temporal stream. In this work, we first investigate whether motion representations are indeed missing in the spatial stream of 3D CNNs. Second, we demonstrate that these motion representations can be improved by distillation, by tuning the spatial stream to predict the outputs of the temporal stream, effectively combining both models into a single stream. Finally, we show that our Distilled 3D Network (D3D) achieves performance on par with two-stream approaches, using only a single model and with no need to compute optical flow.
We present Temporal Aggregation Network (TAN) which decomposes 3D convolutions into spatial and temporal aggregation blocks. By stacking spatial and temporal convolutions repeatedly, TAN forms a deep hierarchical representation for capturing spatio-temporal information in videos. Since we do not apply 3D convolutions in each layer but only apply temporal aggregation blocks once after each spatial downsampling layer in the network, we significantly reduce the model complexity. The use of dilated convolutions at different resolutions of the network helps in aggregating multi-scale spatio-temporal information efficiently. Experiments show that our model is well suited for dense multi-label action recognition, which is a challenging sub-topic of action recognition that requires predicting multiple action labels in each frame. We outperform state-of-the-art methods by 5% and 3% on the Charades and Multi-THUMOS dataset respectively.
Detecting action units (AUs) on human faces is challenging because various AUs make subtle facial appearance change over various regions at different scales. Current works have attempted to recognize AUs by emphasizing important regions. However, the incorporation of expert prior knowledge into region definition remains under-exploited, and current AU detection approaches do not use regional convolutional neural networks (R-CNN) with expert prior knowledge to directly focus on AU-related regions adaptively. By incorporating expert prior knowledge, we propose a novel R-CNN based model named AU R-CNN. The proposed solution offers two main contributions: (1) AU R-CNN directly observes different facial regions, where various AUs are located. Specifically, we define an AU partition rule which encodes the expert prior knowledge into the region definition and RoI-level label definition. This design produces considerably better detection performance than existing approaches. (2) We integrate various dynamic models (including convolutional long short-term memory, two stream network, conditional random field, and temporal action localization network) into AU R-CNN and then investigate and analyze the reason behind the performance of dynamic models. Experiment results demonstrate that \textit{only} static RGB image information and no optical flow-based AU R-CNN surpasses the one fused with dynamic models. AU R-CNN is also superior to traditional CNNs that use the same backbone on varying image resolutions. State-of-the-art recognition performance of AU detection is achieved. The complete network is end-to-end trainable. Experiments on BP4D and DISFA datasets show the effectiveness of our approach. The implementation code is available online.
Existing methods in video action recognition mostly do not distinguish human body from the environment and easily overfit the scenes and objects. In this work, we present a conceptually simple, general and high-performance framework for action recognition in trimmed videos, aiming at person-centric modeling. The method, called Action Machine, takes as inputs the videos cropped by person bounding boxes. It extends the Inflated 3D ConvNet (I3D) by adding a branch for human pose estimation and a 2D CNN for pose-based action recognition, being fast to train and test. Action Machine can benefit from the multi-task training of action recognition and pose estimation, the fusion of predictions from RGB images and poses. On NTU RGB-D, Action Machine achieves the state-of-the-art performance with top-1 accuracies of 97.2% and 94.3% on cross-view and cross-subject respectively. Action Machine also achieves competitive performance on another three smaller action recognition datasets: Northwestern UCLA Multiview Action3D, MSR Daily Activity3D and UTD-MHAD. Code will be made available.
Video action recognition, a critical problem in video understanding, has been gaining increasing attention. To identify actions induced by complex object-object interactions, we need to consider not only spatial relations among objects in a single frame, but also temporal relations among different or the same objects across multiple frames. However, existing approaches that model video representations and non-local features are either incapable of explicitly modeling relations at the object-object level or unable to handle streaming videos. In this paper, we propose a novel dynamic hidden graph module to model complex object-object interactions in videos, of which two instantiations are considered: a visual graph that captures appearance/motion changes among objects and a location graph that captures relative spatiotemporal position changes among objects. Additionally, the proposed graph module allows us to process streaming videos, setting it apart from existing methods. Experimental results on benchmark datasets, Something-Something and ActivityNet, show the competitive performance of our method.
In this paper, we attempt to advance the research work done in human action recognition to a rather specialized application namely Indian Classical Dance (ICD) classification. The variation in such dance forms in terms of hand and body postures, facial expressions or emotions and head orientation makes pose estimation an extremely challenging task. To circumvent this problem, we construct a pose-oblivious shape signature which is fed to a sequence learning framework. The pose signature representation is done in two-fold process. First, we represent person-pose in first frame of a dance video using symmetric Spatial Transformer Networks (STN) to extract good person object proposals and CNN-based parallel single person pose estimator (SPPE). Next, the pose basis are converted to pose flows by assigning a similarity score between successive poses followed by non-maximal suppression. Instead of feeding a simple chain of joints in the sequence learner which generally hinders the network performance we constitute a feature vector of the normalized distance vectors, flow, angles between anchor joints which captures the adjacency configuration in the skeletal pattern. Thus, the kinematic relationship amongst the body joints across the frames using pose estimation helps in better establishing the spatio-temporal dependencies. We present an exhaustive empirical evaluation of state-of-the-art deep network based methods for dance classification on ICD dataset.
One significant problem of deep-learning based human action recognition is that it can be easily misled by the presence of irrelevant objects or backgrounds. Existing methods commonly address this problem by employing bounding boxes on the target humans as part of the input, in both training and testing stages. This requirement of bounding boxes as part of the input is needed to enable the methods to ignore irrelevant contexts and extract only human features. However, we consider this solution is inefficient, since the bounding boxes might not be available. Hence, instead of using a person bounding box as an input, we introduce a human-mask loss to automatically guide the activations of the feature maps to the target human who is performing the action, and hence suppress the activations of misleading contexts. We propose a multi-task deep learning method that jointly predicts the human action class and human location heatmap. Extensive experiments demonstrate our approach is more robust compared to the baseline methods under the presence of irrelevant misleading contexts. Our method achieves 94.06\% and 40.65\% (in terms of mAP) on Stanford40 and MPII dataset respectively, which are 3.14\% and 12.6\% relative improvements over the best results reported in the literature, and thus set new state-of-the-art results. Additionally, unlike some existing methods, we eliminate the requirement of using a person bounding box as an input during testing.