Renguang Chen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Renguang Chen has a total of 14 co-authors. Here are the top co-authors by number of shared publications:
- Chen Feng
Department of Information Engineering, Fuzhou Polytechnic, Fuzhou 350108, China
2 shared publications - Zhide Chen
College of Computer and Cyber Security, Fujian Normal University, Fuzhou, China
2 shared publications - Chen Feng
Department of Information Engineering, Fuzhou Polytechnic, Fuzhou 350003, China
1 shared publication - Hui Cui
Monash University, Melbourne, Australia
1 shared publication - Huijie Guo
Affiliation School of Physical Education and Sport Science, Fujian Normal University, Fuzhou, China
1 shared publication - Jiahao Huang
College of Computer and Cyber Security, Fujian Normal University, Fuzhou 350100, China
1 shared publication
Latest Publications
- Open Access
Skeleton-Based Action Quality Assessment with Anomaly-Aware DTW Optimization for Intelligent Sports Education
In intelligent sports education, current action quality assessment (AQA) methods face significant limitations: regression-based methods are heavily dependent on high-quality annotated data, while unsupervised methods lack sufficient accuracy and degrade performance when handling long-duration sequences. To address these challenges, this paper introduces a novel indirect scoring method integrating action anomaly detection with a Quick Action Quality Assessment (QAQA) algorithm. In this method, the proposed anomaly detection module dynamically adjusts action quality scores by identifying and analyzing acceleration outliers between frames, effectively improving the robustness and accuracy of sports AQA. Moreover, the QAQA algorithm utilizes a multi-resolution approach, including coarsening, projection, and refinement, to significantly reduce computational complexity to , alleviating the computational burden typically associated with long sequence analyses. Experimental results demonstrate that our method outperforms traditional methods in execution efficiency and scoring accuracy. The proposed system improves algorithmic performance and effectively contributes to intelligent sports training and education.Open AccessSkeleton-Based Action Quality Assessment with Anomaly-Aware DTW Optimization for Intelligent Sports Education
In intelligent sports education, current action quality assessment (AQA) methods face significant limitations: regression-based methods are heavily dependent on high-quality annotated data, while unsupervised methods lack sufficient accuracy and degrade performance when handling long-duration sequences. To address these challenges, this paper introduces a novel indirect scoring method integrating action anomaly detection with a Quick Action Quality Assessment (QAQA) algorithm. In this method, the proposed anomaly detection module dynamically adjusts action quality scores by identifying and analyzing acceleration outliers between frames, effectively improving the robustness and accuracy of sports AQA. Moreover, the QAQA algorithm utilizes a multi-resolution approach, including coarsening, projection, and refinement, to significantly reduce computational complexity to , alleviating the computational burden typically associated with long sequence analyses. Experimental results demonstrate that our method outperforms traditional methods in execution efficiency and scoring accuracy. The proposed system improves algorithmic performance and effectively contributes to intelligent sports training and education. - Open Access
Multiscale Two-Stream Fusion Network for Benggang Classification in Multi-Source Images
Benggangs, a type of soil erosion widely distributed in the hilly and mountainous regions of South China, pose significant challenges to land management and ecological conservation. Accurate identification and assessment of their location and scale are essential for effective Benggang control. With advancements in technology, deep learning has emerged as a critical tool for Benggang classification. However, selecting suitable feature extraction and fusion methods for multi-source image data remains a significant challenge. This study proposes a Benggang classification method based on multiscale features and a two-stream fusion network (MS-TSFN). Key features of targeted Benggang areas, such as slope, aspect, curvature, hill shade, and edge, were extracted from Digital Orthophotography Map (DOM) and Digital Surface Model (DSM) data collected by drones. The two-stream fusion network, with ResNeSt as the backbone, extracted multiscale features from multi-source images and an attention-based feature fusion block was developed to explore complementary associations among features and achieve deep fusion of information across data types. A decision fusion block was employed for global prediction to classify areas as Benggang or non-Benggang. Experimental comparisons of different data inputs and network models revealed that the proposed method outperformed current state-of-the-art approaches in extracting spatial features and textures of Benggangs. The best results were obtained using a combination of DOM data, Canny edge detection, and DSM features in multi-source images. Specifically, the proposed model achieved an accuracy of 92.76%, a precision of 85.00%, a recall of 77.27%, and an F1-score of 0.8059, demonstrating its adaptability and high identification accuracy under complex terrain conditions.
Open AccessMultiscale Two-Stream Fusion Network for Benggang Classification in Multi-Source Images
Benggangs, a type of soil erosion widely distributed in the hilly and mountainous regions of South China, pose significant challenges to land management and ecological conservation. Accurate identification and assessment of their location and scale are essential for effective Benggang control. With advancements in technology, deep learning has emerged as a critical tool for Benggang classification. However, selecting suitable feature extraction and fusion methods for multi-source image data remains a significant challenge. This study proposes a Benggang classification method based on multiscale features and a two-stream fusion network (MS-TSFN). Key features of targeted Benggang areas, such as slope, aspect, curvature, hill shade, and edge, were extracted from Digital Orthophotography Map (DOM) and Digital Surface Model (DSM) data collected by drones. The two-stream fusion network, with ResNeSt as the backbone, extracted multiscale features from multi-source images and an attention-based feature fusion block was developed to explore complementary associations among features and achieve deep fusion of information across data types. A decision fusion block was employed for global prediction to classify areas as Benggang or non-Benggang. Experimental comparisons of different data inputs and network models revealed that the proposed method outperformed current state-of-the-art approaches in extracting spatial features and textures of Benggangs. The best results were obtained using a combination of DOM data, Canny edge detection, and DSM features in multi-source images. Specifically, the proposed model achieved an accuracy of 92.76%, a precision of 85.00%, a recall of 77.27%, and an F1-score of 0.8059, demonstrating its adaptability and high identification accuracy under complex terrain conditions.
- Open Access
- 12 April 2024
- Published by MDPI AG in Mathematics
- Vol. 12 (8) , 1162
Automatic Evaluation Method for Functional Movement Screening Based on a Dual-Stream Network and Feature Fusion
Functional Movement Screening (FMS) is a movement pattern quality assessment system used to assess basic movement capabilities such as flexibility, stability, and pliability. Movement impairments and abnormal postures can be identified through peculiar movements and postures of the body. The reliability, validity, and accuracy of functional movement screening are difficult to test due to the subjective nature of the assessment. In this sense, this paper presents an automatic evaluation method for functional movement screening based on a dual-stream network and feature fusion. First, the RAFT algorithm is used to estimate the optical flow of a video, generating a set of optical flow images to represent the motion between consecutive frames. By inputting optical flow images and original video frames separately into the I3D model, it can better capture spatiotemporal features compared to the single-stream method. Meanwhile, this paper introduces a simple but effective attention fusion method that combines features extracted from optical flow with the original frames, enabling the network to focus on the most relevant parts of the input data, thereby improving prediction accuracy. The prediction of the four categories of FMS results was performed. It produced better correlation results compared to other more complex fusion protocols, with an accuracy improvement of 3% over the best-performing fusion method. Tests on public datasets showed that the evaluation metrics of the method proposed in this paper were the most advanced, with an accuracy improvement of approximately 4% compared to the currently superior methods. The use of deep learning methods makes it more objective and reliable to identify human movement impairments and abnormal postures.
- 12 April 2024
- Published by MDPI AG in Mathematics
- Vol. 12 (8) , 1162
Open AccessAutomatic Evaluation Method for Functional Movement Screening Based on a Dual-Stream Network and Feature Fusion
Functional Movement Screening (FMS) is a movement pattern quality assessment system used to assess basic movement capabilities such as flexibility, stability, and pliability. Movement impairments and abnormal postures can be identified through peculiar movements and postures of the body. The reliability, validity, and accuracy of functional movement screening are difficult to test due to the subjective nature of the assessment. In this sense, this paper presents an automatic evaluation method for functional movement screening based on a dual-stream network and feature fusion. First, the RAFT algorithm is used to estimate the optical flow of a video, generating a set of optical flow images to represent the motion between consecutive frames. By inputting optical flow images and original video frames separately into the I3D model, it can better capture spatiotemporal features compared to the single-stream method. Meanwhile, this paper introduces a simple but effective attention fusion method that combines features extracted from optical flow with the original frames, enabling the network to focus on the most relevant parts of the input data, thereby improving prediction accuracy. The prediction of the four categories of FMS results was performed. It produced better correlation results compared to other more complex fusion protocols, with an accuracy improvement of 3% over the best-performing fusion method. Tests on public datasets showed that the evaluation metrics of the method proposed in this paper were the most advanced, with an accuracy improvement of approximately 4% compared to the currently superior methods. The use of deep learning methods makes it more objective and reliable to identify human movement impairments and abnormal postures.