An Improved High-Resolution Network-Based Method for Yoga-Pose Estimation
In this paper, SEPAM_HRNet, a high-resolution pose-estimation model that incorporates the squeeze-and-excitation and pixel-attention-mask (SEPAM) module is proposed. Feature pyramid extraction, channel attention, and pixel-attention masks are integrated into the SEPAM module, resulting in improved m...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-08-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/13/15/8912 |
| _version_ | 1797586996242153472 |
|---|---|
| author | Jianrong Li Dandan Zhang Lei Shi Ting Ke Chuanlei Zhang |
| author_facet | Jianrong Li Dandan Zhang Lei Shi Ting Ke Chuanlei Zhang |
| author_sort | Jianrong Li |
| collection | DOAJ |
| description | In this paper, SEPAM_HRNet, a high-resolution pose-estimation model that incorporates the squeeze-and-excitation and pixel-attention-mask (SEPAM) module is proposed. Feature pyramid extraction, channel attention, and pixel-attention masks are integrated into the SEPAM module, resulting in improved model performance. The construction of the model involves replacing ordinary convolutions with the plug-and-play SEPAM module, which leads to the creation of the SEPAMneck module and SEPAMblock module. To evaluate the model’s performance, the YOGA2022 human yoga poses teaching dataset is presented. This dataset comprises 15,350 images that capture ten basic yoga pose types—Warrior I Pose, Warrior II Pose, Bridge Pose, Downward Dog Pose, Flat Pose, Inclined Plank Pose, Seated Pose, Triangle Pose, Phantom Chair Pose, and Goddess Pose—with a total of five participants. The YOGA2022 dataset serves as a benchmark for evaluating the accuracy of the human pose-estimation model. The experimental results demonstrated that the SEPAM_HRNet model achieved improved accuracy in predicting human keypoints on both the common objects in context (COCO) calibration set and the YOGA2022 calibration set, compared to other state-of-the-art human pose-estimation models with the same image resolution and environment configuration. These findings emphasize the superior performance of the SEPAM_HRNet model. |
| first_indexed | 2024-03-11T00:31:00Z |
| format | Article |
| id | doaj.art-16082164a1fb498bba897efbe4d7e41b |
| institution | Directory Open Access Journal |
| issn | 2076-3417 |
| language | English |
| last_indexed | 2024-03-11T00:31:00Z |
| publishDate | 2023-08-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Applied Sciences |
| spelling | doaj.art-16082164a1fb498bba897efbe4d7e41b2023-11-18T22:39:13ZengMDPI AGApplied Sciences2076-34172023-08-011315891210.3390/app13158912An Improved High-Resolution Network-Based Method for Yoga-Pose EstimationJianrong Li0Dandan Zhang1Lei Shi2Ting Ke3Chuanlei Zhang4College of Artificial Intelligence, Tianjin University of Science and Technology, Tianjin 300453, ChinaCollege of Artificial Intelligence, Tianjin University of Science and Technology, Tianjin 300453, ChinaCollege of Artificial Intelligence, Tianjin University of Science and Technology, Tianjin 300453, ChinaCollege of Artificial Intelligence, Tianjin University of Science and Technology, Tianjin 300453, ChinaCollege of Artificial Intelligence, Tianjin University of Science and Technology, Tianjin 300453, ChinaIn this paper, SEPAM_HRNet, a high-resolution pose-estimation model that incorporates the squeeze-and-excitation and pixel-attention-mask (SEPAM) module is proposed. Feature pyramid extraction, channel attention, and pixel-attention masks are integrated into the SEPAM module, resulting in improved model performance. The construction of the model involves replacing ordinary convolutions with the plug-and-play SEPAM module, which leads to the creation of the SEPAMneck module and SEPAMblock module. To evaluate the model’s performance, the YOGA2022 human yoga poses teaching dataset is presented. This dataset comprises 15,350 images that capture ten basic yoga pose types—Warrior I Pose, Warrior II Pose, Bridge Pose, Downward Dog Pose, Flat Pose, Inclined Plank Pose, Seated Pose, Triangle Pose, Phantom Chair Pose, and Goddess Pose—with a total of five participants. The YOGA2022 dataset serves as a benchmark for evaluating the accuracy of the human pose-estimation model. The experimental results demonstrated that the SEPAM_HRNet model achieved improved accuracy in predicting human keypoints on both the common objects in context (COCO) calibration set and the YOGA2022 calibration set, compared to other state-of-the-art human pose-estimation models with the same image resolution and environment configuration. These findings emphasize the superior performance of the SEPAM_HRNet model.https://www.mdpi.com/2076-3417/13/15/8912human pose estimationattention mechanismhigh-resolution networksfeature pyramids |
| spellingShingle | Jianrong Li Dandan Zhang Lei Shi Ting Ke Chuanlei Zhang An Improved High-Resolution Network-Based Method for Yoga-Pose Estimation Applied Sciences human pose estimation attention mechanism high-resolution networks feature pyramids |
| title | An Improved High-Resolution Network-Based Method for Yoga-Pose Estimation |
| title_full | An Improved High-Resolution Network-Based Method for Yoga-Pose Estimation |
| title_fullStr | An Improved High-Resolution Network-Based Method for Yoga-Pose Estimation |
| title_full_unstemmed | An Improved High-Resolution Network-Based Method for Yoga-Pose Estimation |
| title_short | An Improved High-Resolution Network-Based Method for Yoga-Pose Estimation |
| title_sort | improved high resolution network based method for yoga pose estimation |
| topic | human pose estimation attention mechanism high-resolution networks feature pyramids |
| url | https://www.mdpi.com/2076-3417/13/15/8912 |
| work_keys_str_mv | AT jianrongli animprovedhighresolutionnetworkbasedmethodforyogaposeestimation AT dandanzhang animprovedhighresolutionnetworkbasedmethodforyogaposeestimation AT leishi animprovedhighresolutionnetworkbasedmethodforyogaposeestimation AT tingke animprovedhighresolutionnetworkbasedmethodforyogaposeestimation AT chuanleizhang animprovedhighresolutionnetworkbasedmethodforyogaposeestimation AT jianrongli improvedhighresolutionnetworkbasedmethodforyogaposeestimation AT dandanzhang improvedhighresolutionnetworkbasedmethodforyogaposeestimation AT leishi improvedhighresolutionnetworkbasedmethodforyogaposeestimation AT tingke improvedhighresolutionnetworkbasedmethodforyogaposeestimation AT chuanleizhang improvedhighresolutionnetworkbasedmethodforyogaposeestimation |