Search

In order to solve the problem that the daily rehabilitation training of cardiovascular patients depends on directing by health care professions in rehabilitation centre, during assessing and correcting the movements system about cardiovascular patients independent rehabilitation training at home, this paper proposed an action assessment method for cardiovascular patients' rehabilitation training based on key points of the human skeleton (ASRT-PHS). Firstly, this paper constructed a dataset for rehabilitation training actions using a camera and data augmentation in accordance with the specified rehabilitation training specification for cardiovascular patients. Secondly, this paper employed a deep learning-based detector and pose estimator to capture human body positions and extract key points of the human skeleton, respectively, and then input the results into a convolutional neural network for action recognition. Thirdly, by calculating joint angle thresholds, joint distance ratio and assessing standard motions, this paper constructed a motion segmentation model based on joint distance ratios and an action assessment model based on action joint angle thresholds. This paper investigated the optimal combination of ASRT-PHS by assessing its performance with various joint angle thresholds and action recognition approaches. The results show that ASRT-PHS achieves an average action recognition, segmentation and assessment accuracy of 92.78%, 77.6% and 87%, respectively. Furthermore, case tests about the true cardiovascular patients show that the average accuracy of the prototype system is 71.3%, which provides a feasible intelligent auxiliary system for patients autonomous rehabilitation training at home. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Signal Processing is the property of Journal of Signal Processing and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Information & Control is the property of Gai Kan Bian Wei Hui and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Journal of Computer Engineering & Applications is the property of Beijing Journal of Computer Engineering & Applications Journal Co Ltd. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Journal of Central China Normal University is the property of Huazhong Normal University and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Journal of Computer Engineering & Applications is the property of Beijing Journal of Computer Engineering & Applications Journal Co Ltd. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Journal of Guangdong University of Technology is the property of Journal of Guangdong University of Technology and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

There are some problems in the existing drill pipe counting methods, such as repeated labor, large counting error, and failure to consider the timing information of actions. In order to solve the above problems, a drill pipe counting method based on an improved multi spatial-temporal graph convolution neural network (MST-GCN) model is proposed. Firstly, the video data of underground drilling is obtained through the mine monitoring camera. The Alphabose algorithm is used to extract the key points of the human body from the image sequence. The human skeleton on a single frame image and the skeleton sequence data on a continuous image sequence are obtained. The skeleton sequence representing human actions is built. Secondly, the MST-GCN model is designed based on the spatial-temporal graph convolution neural network (ST-GCN) model. The far space partition strategy is used to focus on the motion information of the key points that are far away from the skeleton. The squeeze and excitation network (SENet) is used to fuse the original space features and the far space features, so as to effectively identify the action categories on the skeleton sequence. Finally, support vector machine is used to identify the drilling pose on the drilling video to decide whether to save the skeleton sequence. If the sequence length is saved to 150 frames, the MST-GCN model is used to identify the action category. The identification interval of adjacent actions is set according to the actual drilling time, so as to record the number of actions and realize the drill pipe counting. The experimental results show that the recognition accuracy of the MST-GCN model is 91.1% on the self-built data set, which is 6.2%, 19.0% and 4.8% higher than that of ST-GCN, Alphapose-LSTM and NST-GCN, respectively. The loss value of the MST-GCN model converges below 0.2, and the learning capability is stronger. On the drilling videos under the same conditions, the average error values of the MST-GCN model, the artificial method and the improved ResNet method are 0.25, 0.75 and 21 respectively, which shows that the counting effect of the MST-GCN model is better. The average error of MST-GCN model is 9 and the miscount is low in the field application of drilling 1 300 pieces, which can meet the actual requirements.

Copyright of Journal of South China Normal University (Natural Science Edition) / Huanan Shifan Daxue Xuebao (Ziran Kexue Ban) is the property of Journal of South China Normal University (Natural Science Edition) Editorial Office and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of China Sciencepaper is the property of China Sciencepaper and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Journal of Frontiers of Computer Science & Technology is the property of Beijing Journal of Computer Engineering & Applications Journal Co Ltd. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Journal of Test & Measurement Technology is the property of Publishing Center of North University of China and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Journal of South China Normal University (Natural Science Edition) / Huanan Shifan Daxue Xuebao (Ziran Kexue Ban) is the property of Journal of South China Normal University (Natural Science Edition) Editorial Office and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Journal of Mine Automation is the property of Industry & Mine Automation Editorial Department and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Journal of Data Acquisition & Processing / Shu Ju Cai Ji Yu Chu Li is the property of Editorial Department of Journal of Nanjing University of Aeronautics & Astronautics and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Chinese Journal of Liquid Crystal & Displays is the property of Chinese Journal of Liquid Crystal & Displays and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Journal of Beijing University of Technology is the property of Journal of Beijing University of Technology, Editorial Department and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Under the monitoring scenario, due to the shortage of monitoring resources, abnormal pedestrian behavior is prone to miss detection. In view of this problem, a method to identify abnormal human behavior in video surveillance scene is proposed to assist monitoring personnel to find abnormal behavior in time. The human skeleton is extracted from the pedestrians in the image using OpenPose. The graph attention mechanism based on the graph attention network (GAT) is integrated to solve the problem that the graph convolution network has a single way to aggregate the features of joint points. On the basis of the improved graph convolutional network, the spatial temporal graph convolutional networks (ST-GCN) is used to identify the abnormal behavior of pedestrians joint points information. The experimental results show that the recognition algorithm proposed in this paper has an accuracy of 85.48% for the defined behavior, and can accurately identify the abnormal behavior of pedestrians in surveillance video. [ABSTRACT FROM AUTHOR]

Aiming at the problem of missing detection and false detection caused by inaccurate motion feature extraction of existing video key frame extraction algorithms, this paper proposed a reinforcement learning key frame extraction algorithm combining multi-channel feature and attention mechanism. The algorithm extracted the human skeleton joint points from the video sequence through the human posture recognition algorithm firstly. Then it used the S-GCN and ResNet50 network to extract the motion features and static features in the video sequence respectively, and performed a weighted fusion of the two. Finally, it applied the attention mechanism to calculate the importance of the video frame of the feature sequence, and used reinforcement learning to extract and optimize key frames. The experimental results show that the algorithm can solve the problem of missing and false detection in the key frame extraction of motion video. It performs well in the detection of video frames containing key actions, with high accuracy and strong stability. [ABSTRACT FROM AUTHOR]

Copyright of Chinese Journal of Liquid Crystal & Displays is the property of Chinese Journal of Liquid Crystal & Displays and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Action recognition methods based on skeleton data have received extensive attention and research due to their strong adaptability to dynamic environments and complex backgrounds. The application of graph convolutional networks to describe human skeleton to realize human action recognition can achieve good recognition results, but the topological structure of the graph is often manually set, and the structure on all layers and input samples is fixed. Also the graph convolutional networks can only capture the local physical relationship between joints, and miss the correlation of non-physical joints. This paper proposes a new skeleton action recognition based on regional association adaptive graph convolutional network. Through adaptive graph convolution, the structure of the parameterized global graph and the single data graph and model convolution parameters are trained and updated in different layers, increasing the flexibility of the graph structure in the model and the versatility of the model for various data samples. This paper introduces the regional association graph convolution, and the non-physical connection correlation of each joint between data frames is captured by alternating information transfer between joint features and connection features. And it adds the second-order data of the skeleton to supplement the original joint data, merges this two to form a two-stream network to improve the performance of the recognition network. Experiments on the NTU-RGBD large-scale dataset show that the model has a certain improvement in the accuracy of action recognition. [ABSTRACT FROM AUTHOR]

Compared with the traditional RGB-based methods, the skeleton-based action recognition methods have become the main research direction in the field of computer vision in recent years because they are less affected by many factors such as illumination, viewing angle and background complexity. However, the current skeleton-based methods still have some problems such as large parameters, long time-consuming and high computational complexity, which makes it complicated and difficult to meet the requirements of efficiency and accuracy simultaneously. To address these issues, a lightweight graph convolution network using multi-modal data fusion is proposed. Firstly, the multi-modal information flow data are fused by multi-modal data fusion method. Secondly, the spatial and temporal information of human joints are extracted using spatial and temporal modules respectively. Finally, the classification results are obtained through the fully connected layer. Experimental results conducted on the two commonly used datasets including NTU60 RGB+D and NTU120 RGB+D demonstrate that the proposed network outperforms some mainstream methods in the last two years in both recognition accuracy and efficiency, thus verifying that the network has excellent performance in terms of accuracy, while considering time efficiency and computational cost.

The traditional RGB-based action recognition method is vulnerable to background factors such as illumination and occlusion, which makes the recognition accuracy and real-time performance low. In order to solve the above problems, a 2 D skeleton action recognition method incorporating a hierarchical attention mechanism was proposed. Firstly, the human skeleton joints in the video were extracted using the human pose estimation algorithm OpenPose. Secondly, the missing and abnormal points in the extracted skeleton data were pre-processed by the mean-completion method and exponential smoothing method. Finally, a network model based on CNN-LSTM fused with hierarchical attention mechanism, CNN-HALSTM, was constructed to realize action classification. The results show that the recognition accuracy of this method in interactive motion and KTH data sets is 96.73% and 98.35%, and the model parameters are significantly reduced, which has the advantage of real-time performance and is better than other methods of the same type. [ABSTRACT FROM AUTHOR]

Aiming at the problems that existing human skeleton action recognition methods couldn't explore sufficient human body information and extract sufficient temporal feature, this paper proposed a model based on posture transformation module and posture fusion module (PTF-SGN), which realized the utilization of the key spatio-temporal information in skeleton diagram. Firstly, by preprocessing the skeleton diagram, the model mined the displacement information of limbs and joints, and extracted the features. Then it used the posture transformation module to obtain the posture adjustment factors from the skeleton image data in an unsupervised learning manner, and adaptively adjusted the body posture to enhance the robustness of the model in different environments. Secondly, it proposed a posture fusion module based on the time attention mechanism, which learned the short-term features and the long-term features, and fused the time characteristics of long and short moments to strengthen the characterization ability of time characteristics. Finally, it extracted the global spatio-temporal feature of the skeleton feature to input into the classification network to obtain the action recognition result. The experimental results on the two 3D skeleton datasets of NTU60 RGB+D and NTU120 RGB+D and the two 2D skeleton datasets of Penn-Action and HARPET show that PTF-SGN model can effectively recognize actions of skeleton time series data. [ABSTRACT FROM AUTHOR]

In order to effectively describe the spatio-temporal information of human action, this paper proposed skeletal features by Hough transform to construct the motion representation. Specifically, under the OpenPose framework, the method obtained the key points of human skeletons in video frames, and built the skeleton joints, then converted the trajectories of skeleton joints into the point trajectories in Hough space. After that, FV（Fisher vector） based approach coded the angle and trajectory features, and then the concatenate method combined the coded features to input the linear SVM classifier. On the classical public datasets KTH, Weizmann, KARD and Drone-Action. Experimental results show that the Hough transform promotes the robustness of features, and improves the performance of human action recognition. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Signal Processing is the property of Journal of Signal Processing and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Journal of Frontiers of Computer Science & Technology is the property of Beijing Journal of Computer Engineering & Applications Journal Co Ltd. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Based on the advantages of robustness and view-invariant representation, a skeleton-based action recognition algorithm combining geometric features with LSTM network is proposed to further improve the recognition rate and to break the limitation that the inputs of most methods based on deep learning are human joint coordinates. The geometric features based on the distances between joints and selected lines are selected as the input of the network. Then, time-selective LSTM network is introduced to train. Time selection LSTM network has the ability to select the most recognizable time period features. By using this feature, 99.36% and 99.20% recognition rates are achieved on SBU Interaction dataset and UT Kinect dataset, respectively. The experimental results show that the method is effective for human skeleton-based action recognition. [ABSTRACT FROM AUTHOR]

In order to imp rove the modeling efficiency of 3D human skeleton model and simplify the interaction rules, this pa· per presented a deep learning based 3 D skeleton model reconstruction method from human motion sketch. Firstly, this method rendered the 3D skeleton models into 2D images to establish the dimension mapping and then used the image classification method to recognize motion from sketch and further to realize 3 D skeleton model reconstruction according to the dimension mapping between 2 D and 3D. This method built the image classification model base on CNN and used a shallow convolutional network as the training unit in the experiment. This method also used hierarchical classification and blocking training scheme to accelerate the convergence time of network to improve training efficiency. Finally, the experiment results verify the feasibility and effectiveness of this method. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Image & Graphics is the property of Editorial Office of Journal of Image & Graphics and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Aiming at fingertip detection in a complex environment, from the two aspects of hand image segmentation and fingertip detection, a fingertip detection method based on human skeleton and depth image information is proposed. Firstly, Kinect is used to obtain the human skeleton and depth image information, the objective user is locked by the human skeleton information, and then the hand joint position of the locked user is used to extract the hand area image from the depth image. Secondly, the skeleton endpoints are searched from the hand skeleton, and the local optimal search method is proposed to optimize the results of contour convex hull computation. Finally, the fingertip position is found by combining the feature of hand contour. The experimental results show that the proposed method has good detection effect, meets the real-time requirements and can realize robust detection in a complex environment. [ABSTRACT FROM AUTHOR]