Action Recognition Method Research Articles

Retracted: Dance-Specific Action Recognition Method Based on Double-Stream CNN in Complex Environment.

[This retracts the article DOI: 10.1155/2022/9327277.].

Tell me what you see: A zero-shot action recognition method based on natural language descriptions

Tell me what you see: A zero-shot action recognition method based on natural language descriptions

Human activity recognition method using joint deep learning and acceleration signal

Many studies have been conducted on human activity recognition (HAR) in the last decade. Accordingly, deep learning (DL) algorithms have been given more attention in terms of classification of human daily activities. Deep neural networks (DNNs) compute and extract complex features on voluminous data through some hidden layers that require large memory and powerful graphics processing units (GPUs). So, this study proposes a new joint learning (JL) approach to classify human activities using inertial sensors. To this end, a large complex donor model based on a convolutional neural network (CNN) is used to transfer knowledge to a smaller model based on CNN referred to as the acceptor model. The acceptor model can be deployed on mobile devices and low-power hardware due to decreased computing costs and memory consumption. The wireless sensor data mining (WISDM) dataset is used to test the proposed model. According to the experimental results, the HAR system based on the JL algorithm outperforms than other methods.

Weakly Supervised Temporal Convolutional Networks for Fine-Grained Surgical Activity Recognition.

Automatic recognition of fine-grained surgical activities, called steps, is a challenging but crucial task for intelligent intra-operative computer assistance. The development of current vision-based activity recognition methods relies heavily on a high volume of manually annotated data. This data is difficult and time-consuming to generate and requires domain-specific knowledge. In this work, we propose to use coarser and easier-to-annotate activity labels, namely phases, as weak supervision to learn step recognition with fewer step annotated videos. We introduce a step-phase dependency loss to exploit the weak supervision signal. We then employ a Single-Stage Temporal Convolutional Network (SS-TCN) with a ResNet-50 backbone, trained in an end-to-end fashion from weakly annotated videos, for temporal activity segmentation and recognition. We extensively evaluate and show the effectiveness of the proposed method on a large video dataset consisting of 40 laparoscopic gastric bypass procedures and the public benchmark CATARACTS containing 50 cataract surgeries.

SKELTER: unsupervised skeleton action denoising and recognition using transformers

Unsupervised Human Action Recognition (U-HAR) methods currently leverage large-scale datasets of human poses to solve this challenging problem. As most of the approaches are dedicated to reaching the best recognition accuracies, no attention has been put into analyzing the resilience of such methods given perturbed data, a likely occurrence in real in-the-wild testing scenarios. Our first contribution is to systematically validate the decrease in performance of current U-HAR state-of-the-art using perturbed or altered data (e.g., obtained by removing some skeletal joints, rotating the entire pose, and injecting geometrical aberrations). Then, we propose a novel framework based on a transformer encoder–decoder with remarkable de-noising capabilities to counter such perturbations effectively. Moreover, we also present additional losses to have robust representations against rotation variances and provide temporal motion consistency. Our model, SKELTER, shows limited drops in performance when skeleton noise is present compared with previous approaches, favoring its use in challenging in-the-wild settings.

An effective PoseC3D model for typical action recognition of dairy cows based on skeleton features

The precise recognition of dairy cows' motion behaviors is crucial for the intelligent assessment of their health status. Previous researchers have utilized many methods to detect cow’s motion behaviors, but the accuracy and robustness of these methods remains to be improved. Skeleton-based action recognition methods have reported high accuracy and robustness, but these methods have rarely been applied to the detection of cow’s motion behavior. This research adopted a behavior recognition method based on skeleton features of dairy cows. First, the YOLOX was modified to YOLOX-Pose for end-to-end skeleton extraction. According to the extracted information, the PoseC3D based on 3D Convolutional Neural Network (CNN) was used to predict motion behaviors. In the process of model training, PoseC3D utilized a dataset consisting of 400 videos encompassing actions of lying, standing, walking, and lameness, while YOLOX-Pose was trained using a dataset comprising 1800 images. The results showed that, for skeleton extraction, YOLOX-Pose achieved an AP of 96.3% and detection speed of 33.1fps. For the recognition of the four actions using PoseC3D, the Top-1 accuracy was 92.6%. The results showed that the YOLOX-Pose was a lightweight and high-quality pose estimator, and PoseC3D, with extracted 2D skeleton information, could outperform Graph Convolutional Network (GCN) in action detection. The combination of the YOLOX-Pose and PoseC3D could be used to detect the four typical motion behaviors effectively.

Retracted: Study about Football Action Recognition Method Based on Deep Learning and Improved Dynamic Time Warping Algorithm

Retracted: Study about Football Action Recognition Method Based on Deep Learning and Improved Dynamic Time Warping Algorithm

3D-unified spatial-temporal graph for group activity recognition

Early group activity recognition is typically conducted in a 2D scenario. This paper proposes a group activity recognition method in a 3D space. In practical applications, differences in camera angle and mutual shielding during the peak flow of people will affect the extraction of individual characteristics and the calculation of individual relationships. Therefore, a method using 3D relationships is proposed. We first detect the 3D backbone of pedestrians and calculate the movement relationship of individuals in the group based on it. Second, we calculate the 3D-unified spatial–temporal graphs (3D-USTG) of the group and the individuals in the group using global-G3D (G-G3D). The experimental results demonstrate that the proposed method can effectively solve the problems caused by occlusion and shooting angle compared with a 2D method, which indicates that the proposed method can achieve better results when applied to intersections of different specifications. Additionally, the proposed method achieves strong performance metrics with two publicly available datasets for group activity recognition.

Wearable sensor-based human activity recognition with ensemble learning: a comparison study

<p>The spectacular growth of wearable sensors has provided a key contribution to the field of human activity recognition. Due to its effective and versatile usage and application in various fields such as smart homes and medical areas, human activity recognition has always been an appealing research topic in artificial intelligence. From this perspective, there are a lot of existing works that make use of accelerometer and gyroscope sensor data for recognizing human activities. This paper presents a comparative study of ensemble learning methods for human activity recognition. The methods include random forest, adaptive boosting, gradient boosting, extreme gradient boosting, and light gradient boosting machine (LightGBM). Among the ensemble learning methods in comparison, light gradient boosting machine and random forest demonstrate the best performance. The experimental results revealed that light gradient boosting machine yields the highest accuracy of 94.50% on UCI-HAR dataset and 100% on single accelerometer dataset while random forest records the highest accuracy of 93.41% on motion sense dataset.</p>

A Human Activity Recognition Method Based on Lightweight Feature Extraction Combined With Pruned and Quantized CNN for Wearable Device

Human Activity Recognition (HAR) is becoming an essential part of human life care. Existing HAR methods are usually developed using a two-level approach, wherein a first-level Machine Learning (ML) classifier is employed to distinguish the static and dynamic activities, followed by a second-level classifier to identify the specific activity. These approaches are not suitable for wearable devices, due to the high computational and memory consumption. Our rigorous analysis of various HAR datasets opens up a new possibility that static or dynamic activities can be discriminated against through a simple statistical technique. Therefore, we propose to utilize a statistical feature extraction technique to replace the first-level ML classifier, thus achieving more lightweight computation. Next, we employ Random Forest (RF) and Convolutional Neural Networks (CNN) to classify the specific activities, achieving higher accuracy compared to the state-of-the-art results. We further reduce the computation and memory consumption of the above combined approach by applying pruning and quantizing techniques to CNN (PQ-CNN). Experimental results show the proposed lightweight HAR method achieved an F1 score of 0.9417 and 0.9438 for unbalanced and balanced datasets, respectively. On top of lightweight and accuracy, the proposed HAR method is practical for wearable devices by using a single accelerometer.

Human Action Representation Learning Using an Attention-Driven Residual 3DCNN Network

The recognition of human activities using vision-based techniques has become a crucial research field in video analytics. Over the last decade, there have been numerous advancements in deep learning algorithms aimed at accurately detecting complex human actions in video streams. While these algorithms have demonstrated impressive performance in activity recognition, they often exhibit a bias towards either model performance or computational efficiency. This biased trade-off between robustness and efficiency poses challenges when addressing complex human activity recognition problems. To address this issue, this paper presents a computationally efficient yet robust approach, exploiting saliency-aware spatial and temporal features for human action recognition in videos. To achieve effective representation of human actions, we propose an efficient approach called the dual-attentional Residual 3D Convolutional Neural Network (DA-R3DCNN). Our proposed method utilizes a unified channel-spatial attention mechanism, allowing it to efficiently extract significant human-centric features from video frames. By combining dual channel-spatial attention layers with residual 3D convolution layers, the network becomes more discerning in capturing spatial receptive fields containing objects within the feature maps. To assess the effectiveness and robustness of our proposed method, we have conducted extensive experiments on four well-established benchmark datasets for human action recognition. The quantitative results obtained validate the efficiency of our method, showcasing significant improvements in accuracy of up to 11% as compared to state-of-the-art human action recognition methods. Additionally, our evaluation of inference time reveals that the proposed method achieves up to a 74× improvement in frames per second (FPS) compared to existing approaches, thus showing the suitability and effectiveness of the proposed DA-R3DCNN for real-time human activity recognition.

Enhanced Spatial Stream of Two-Stream Network Using Optical Flow for Human Action Recognition

Introduction: Convolutional neural networks (CNNs) have maintained their dominance in deep learning methods for human action recognition (HAR) and other computer vision tasks. However, the need for a large amount of training data always restricts the performance of CNNs. Method: This paper is inspired by the two-stream network, where a CNN is deployed to train the network by using the spatial and temporal aspects of an activity, thus exploiting the strengths of both networks to achieve better accuracy. Contributions: Our contribution is twofold: first, we deploy an enhanced spatial stream, and it is demonstrated that models pre-trained on a larger dataset, when used in the spatial stream, yield good performance instead of training the entire model from scratch. Second, a dataset augmentation technique is presented to minimize overfitting of CNNs, where we increase the dataset size by performing various transformations on the images such as rotation and flipping, etc. Results: UCF101 is a standard benchmark dataset for action videos, and our architecture has been trained and validated on it. Compared with the other two-stream networks, our results outperformed them in terms of accuracy.

Advancing human action recognition: A hybrid approach using attention-based LSTM and 3D CNN

In this paper, we propose a novel approach to video action recognition that integrates a modified and optimized 3D Convolutional Neural Network, a Long Short-Term Memory network, and attention mechanisms. This synergy enhances the overall performance, offering an advantage over existing methods in managing the intricacies of real-world scenarios. The uniqueness of our approach lies in its capacity to capture both spatial and temporal information from video sequences and the incorporation of an attention mechanism that selectively emphasizes key areas within the sequences, thereby enhancing recognition accuracy. The model is particularly tailored to handle complex scenarios, such as those with multiple actors or objects, or instances of occlusion. It effectively addresses the subjectivity and variability inherent in action annotations within datasets. We also apply an array of preprocessing techniques to further optimize model performance. Through rigorous experimental evaluations on benchmark datasets, namely UCF101 and HMDB51, we demonstrate that our proposed approach significantly outperforms existing state-of-the-art methods in action recognition. These results underscore the potential of our approach for further advancements in video action recognition research.

Two-Person Graph Convolutional Network for Skeleton-Based Human Interaction Recognition

Graph convolutional networks (GCNs) have been the predominant methods in skeleton-based human action recognition, including human-human interaction recognition. However, when dealing with interaction sequences, current GCN-based methods simply split the two-person skeleton into two discrete graphs and perform graph convolution separately as done for single-person action classification. Such operations ignore rich interactive information and hinder effective spatial inter-body relationship modeling. To overcome the above shortcoming, we introduce a novel unified two-person graph to represent inter-body and intra-body correlations between joints. Experimental results show accuracy improvements in recognizing both interactions and individual actions when utilizing the proposed two-person graph topology. In addition, several graph labeling strategies are designed to supervise the model to learn discriminant spatial-temporal interactive features. Finally, we propose a two-person graph convolutional network (2P-GCN). Our model outperforms state-of-the-art methods on four benchmarks of three interaction datasets: SBU, interaction subsets of NTU-RGB+D and NTU-RGB+D 120.

CHAN: Skeleton based action recognition by multi‐level feature learning

AbstractSkeleton‐based action recognition has been continuously and intensively studied. However, dynamic 3D skeleton data are difficult to be popularized in practical applications due to the restricted data acquisition conditions. Although the action recognition method based on 2D pose information extracted from RGB video can effectively avoid the influence of complex background, it is susceptible to factors such as video jitter and joint overlap. To reduce the interference of the aforementioned factors, we use two‐dimensional skeletal joint coordinate modal information to represent the changes in human body posture. First, we use a target detector and pose estimation algorithm to obtain the joint coordinates of each frame sample from RGB video. Then the feature extraction network is combined to perform multi‐level feature learning to establish correspondence between actions and corresponding multi‐level features. Finally, the hierarchical attention mechanism is introduced to design the model named CHAN. By calculating the association between elements, the weight of the action classification is redistributed. Extensive experiments on three datasets demonstrate the effectiveness of our proposed method.

Spatial Hard Attention Modeling via Deep Reinforcement Learning for Skeleton-Based Human Activity Recognition

Deep learning-based algorithms have been very successful in skeleton-based human activity recognition. Skeleton data contains 2-D or 3-D coordinates of human body joints. The main focus of most of the existing skeleton-based activity recognition methods is on designing new deep architectures to learn discriminative features, where all body joints are considered equally important in recognition. However, the importance of joints varies as an activity proceeds within a video and across different activities. In this work, we hypothesize that selecting relevant joints, prior to recognition, can enhance performance of the existing deep learning-based recognition models. We propose a spatial hard attention finding method that aims to remove the uninformative and/or misleading joints at each frame. We formulate the joint selection problem as a Markov decision process and employ deep reinforcement learning to train the proposed spatial-attention-aware agent. No extra labels are needed for the agent’s training. The agent takes a sequence of features extracted from skeleton video as input and outputs a sequence of probabilities for joints. The proposed method can be considered as a general framework that can be integrated with the existing skeleton-based activity recognition methods for performance improvement purposes. We obtain very competitive activity recognition results on three commonly used human activity recognition datasets.

Action Recognition via Adaptive Semi-Supervised Feature Analysis

This study presents a new semi-supervised action recognition method via adaptive feature analysis. We assume that action videos can be regarded as data points in embedding manifold subspace, and their matching problem can be quantified through a specific Grassmannian kernel function while integrating feature correlation exploration and data similarity measurement into a joint framework. By maximizing the intra-class compactness based on labeled data, our algorithm can learn multiple features and leverage unlabeled data to enhance recognition. We introduce the Grassmannian kernels and the Projected Barzilai–Borwein (PBB) method to train a subspace projection matrix as a classifier. Experiment results show our method has outperformed the compared approaches when a few labeled training samples are available.

Novel Motion Patterns Matter for Practical Skeleton-Based Action Recognition

Most skeleton-based action recognition methods assume that the same type of action samples in the training set and the test set share similar motion patterns. However, action samples in real scenarios usually contain novel motion patterns which are not involved in the training set. As it is laborious to collect sufficient training samples to enumerate various types of novel motion patterns, this paper presents a practical skeleton-based action recognition task where the training set contains common motion patterns of action samples and the test set contains action samples that suffer from novel motion patterns. For this task, we present a Mask Graph Convolutional Network (Mask-GCN) to focus on learning action-specific skeleton joints that mainly convey action information meanwhile masking action-agnostic skeleton joints that convey rare action information and suffer more from novel motion patterns. Specifically, we design a policy network to learn layer-wise body masks to construct masked adjacency matrices, which guide a GCN-based backbone to learn stable yet informative action features from dynamic graph structure. Extensive experiments on our newly collected dataset verify that Mask-GCN outperforms most GCN-based methods when testing with various novel motion patterns.

Human Activity Recognition Using Cascaded Dual Attention CNN and Bi-Directional GRU Framework.

Vision-based human activity recognition (HAR) has emerged as one of the essential research areas in video analytics. Over the last decade, numerous advanced deep learning algorithms have been introduced to recognize complex human actions from video streams. These deep learning algorithms have shown impressive performance for the video analytics task. However, these newly introduced methods either exclusively focus on model performance or the effectiveness of these models in terms of computational efficiency, resulting in a biased trade-off between robustness and computational efficiency in their proposed methods to deal with challenging HAR problem. To enhance both the accuracy and computational efficiency, this paper presents a computationally efficient yet generic spatial-temporal cascaded framework that exploits the deep discriminative spatial and temporal features for HAR. For efficient representation of human actions, we propose an efficient dual attentional convolutional neural network (DA-CNN) architecture that leverages a unified channel-spatial attention mechanism to extract human-centric salient features in video frames. The dual channel-spatial attention layers together with the convolutional layers learn to be more selective in the spatial receptive fields having objects within the feature maps. The extracted discriminative salient features are then forwarded to a stacked bi-directional gated recurrent unit (Bi-GRU) for long-term temporal modeling and recognition of human actions using both forward and backward pass gradient learning. Extensive experiments are conducted on three publicly available human action datasets, where the obtained results verify the effectiveness of our proposed framework (DA-CNN+Bi-GRU) over the state-of-the-art methods in terms of model accuracy and inference runtime across each dataset. Experimental results show that the DA-CNN+Bi-GRU framework attains an improvement in execution time up to 167× in terms of frames per second as compared to most of the contemporary action-recognition methods.

Hang-Time HAR: A Benchmark Dataset for Basketball Activity Recognition Using Wrist-Worn Inertial Sensors.

We present a benchmark dataset for evaluating physical human activity recognition methods from wrist-worn sensors, for the specific setting of basketball training, drills, and games. Basketball activities lend themselves well for measurement by wrist-worn inertial sensors, and systems that are able to detect such sport-relevant activities could be used in applications of game analysis, guided training, and personal physical activity tracking. The dataset was recorded from two teams in separate countries (USA and Germany) with a total of 24 players who wore an inertial sensor on their wrist, during both a repetitive basketball training session and a game. Particular features of this dataset include an inherent variance through cultural differences in game rules and styles as the data was recorded in two countries, as well as different sport skill levels since the participants were heterogeneous in terms of prior basketball experience. We illustrate the dataset's features in several time-series analyses and report on a baseline classification performance study with two state-of-the-art deep learning architectures.