Action Recognition Scheme Based on Skeleton Representation With DS-LSTM Network

Jiang, Xinghao; Xu, Ke; Sun, Tanfeng

doi:10.1109/tcsvt.2019.2914137

Cited by 65 publications

(29 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The collected image data is retrieved by the FPGA and stored in the system through the latent memory. DDR2 SDRAM is used to store the data and finally stored in the NAND memory [18]. The stored data will eventually be uploaded to the computer for processing or storage.…”

Section: B Image Acquisition System Constructionmentioning

confidence: 99%

Aided Image Acquisition System for Aerobics Training Based on Motion Recognition Technology

Hai-ying

Liu

2024

IEEE Access

View full text Add to dashboard Cite

The image acquisition system is the main part of acquiring image information, and its performance largely determines the accuracy and difficulty of subsequent planning. The main purpose of this article is to design an image acquisition system for aerobics training based on motion recognition technology. This article mainly introduces the teaching experiments based on the design of aerobics courses. Through the empirical investigation and analysis of the impact of relevant experimental data on aviation training courses, attempts to establish the connection between aerobics social adaptability and various dimensions The cultivation of student' s social adaptability. In this paper, the GMM algorithm is mainly used to distinguish the rest time when the action occurs, and the subsequent rest period is used as the basis for segmenting multiple events in the action sequence. Finally, the characteristics of the action coding mapping of each event are derived, and the support vector machine is used to complete the energy recognition process of the existence of a single energy. The experimental results of this paper show that the designed embedded image acquisition system has high integration and stability, the acquired image resolution is 640x480, and the wireless transmission rate is 5MbPs, which has wide application prospects.

show abstract

Section: B Image Acquisition System Constructionmentioning

confidence: 99%

Aided Image Acquisition System for Aerobics Training Based on Motion Recognition Technology

Hai-ying

Liu

2024

IEEE Access

View full text Add to dashboard Cite

show abstract

“…Therefore, the spatial relations of joints and body structure are ignored. DS-LSTM [57] trained a three-layer denoising sparse LSTM structure to capture the temporal connections in the skeleton sequence. Nevertheless, only the information of the previous time step was used with LSTM.…”

Section: Comparisons With the State-of-the-artmentioning

confidence: 99%

“…ST-LSTM [6] 69.20 77.70 Two-stream RNN [15] 71.30 79.50 GCA-LSTM [12] 74.40 82.80 Visualization CNN [35] 76.00 82.60 VA-LSTM [39] 79.40 87.60 Clips+CNN+MTLN [36] 79.57 84.83 ST-GCN [20] 81.50 88.30 3D-GCN [56] 82.60 89.60 DS-LSTM [57] 77.80 87.33 AAM-GCN (ours) 82.73 90.12 HDM05 is a motion capture dataset utilized on skeleton-based works. Compared with other large-scale datasets that have emerged in recent years, it is much smaller.…”

Section: Cross Subject (%) Cross View (%)mentioning

confidence: 99%

Adaptive Attention Memory Graph Convolutional Networks for Skeleton-Based Action Recognition

Liu

Wang

et al. 2021

Sensors

View full text Add to dashboard Cite

Graph Convolutional Networks (GCNs) have attracted a lot of attention and shown remarkable performance for action recognition in recent years. For improving the recognition accuracy, how to build graph structure adaptively, select key frames and extract discriminative features are the key problems of this kind of method. In this work, we propose a novel Adaptive Attention Memory Graph Convolutional Networks (AAM-GCN) for human action recognition using skeleton data. We adopt GCN to adaptively model the spatial configuration of skeletons and employ Gated Recurrent Unit (GRU) to construct an attention-enhanced memory for capturing the temporal feature. With the memory module, our model can not only remember what happened in the past but also employ the information in the future using multi-bidirectional GRU layers. Furthermore, in order to extract discriminative temporal features, the attention mechanism is also employed to select key frames from the skeleton sequence. Extensive experiments on Kinetics, NTU RGB+D and HDM05 datasets show that the proposed network achieves better performance than some state-of-the-art methods.

show abstract

“…Human action recognition is a challenging and attractive research topic in computer vision. It can be applied in various applications, such as video understanding, intelligent surveillance, security, robotics, human–computer interactions, industrial automation, health care, and education [ 1 , 2 , 3 , 4 ]. In spite much research work in this domain, many challenges in action recognition have remained unresolved.…”

Section: Introductionmentioning

confidence: 99%

Student Behavior Recognition System for the Classroom Environment Based on Skeleton Pose Estimation and Person Detection

Lin

Ngo

Dow

et al. 2021

Sensors

View full text Add to dashboard Cite

Human action recognition has attracted considerable research attention in the field of computer vision, especially for classroom environments. However, most relevant studies have focused on one specific behavior of students. Therefore, this paper proposes a student behavior recognition system based on skeleton pose estimation and person detection. First, consecutive frames captured with a classroom camera were used as the input images of the proposed system. Then, skeleton data were collected using the OpenPose framework. An error correction scheme was proposed based on the pose estimation and person detection techniques to decrease incorrect connections in the skeleton data. The preprocessed skeleton data were subsequently used to eliminate several joints that had a weak effect on behavior classification. Second, feature extraction was performed to generate feature vectors that represent human postures. The adopted features included normalized joint locations, joint distances, and bone angles. Finally, behavior classification was conducted to recognize student behaviors. A deep neural network was constructed to classify actions, and the proposed system was able to identify the number of students in a classroom. Moreover, a system prototype was implemented to verify the feasibility of the proposed system. The experimental results indicated that the proposed scheme outperformed the skeleton-based scheme in complex situations. The proposed system had a 15.15% higher average precision and 12.15% higher average recall than the skeleton-based scheme did.

show abstract

Action Recognition Scheme Based on Skeleton Representation With DS-LSTM Network

Cited by 65 publications

References 40 publications

Aided Image Acquisition System for Aerobics Training Based on Motion Recognition Technology

Aided Image Acquisition System for Aerobics Training Based on Motion Recognition Technology

Adaptive Attention Memory Graph Convolutional Networks for Skeleton-Based Action Recognition

Student Behavior Recognition System for the Classroom Environment Based on Skeleton Pose Estimation and Person Detection

Contact Info

Product

Resources

About