A Lightweight 6D Pose Estimation Network Based on Improved Atrous Spatial Pyramid Pooling
It is difficult for lightweight neural networks to produce accurate 6DoF pose estimation effects due to their accuracy being affected by scale changes. To solve this problem, we propose a method with good performance and robustness based on previous research. The enhanced PVNet-based method uses dep...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2024-04-01
|
| Series: | Electronics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2079-9292/13/7/1321 |
| _version_ | 1797212739563683840 |
|---|---|
| author | Fupan Wang Xiaohang Tang Yadong Wu Yinfan Wang Huarong Chen Guijuan Wang Jing Liao |
| author_facet | Fupan Wang Xiaohang Tang Yadong Wu Yinfan Wang Huarong Chen Guijuan Wang Jing Liao |
| author_sort | Fupan Wang |
| collection | DOAJ |
| description | It is difficult for lightweight neural networks to produce accurate 6DoF pose estimation effects due to their accuracy being affected by scale changes. To solve this problem, we propose a method with good performance and robustness based on previous research. The enhanced PVNet-based method uses depth-wise convolution to build a lightweight network. In addition, coordinate attention and atrous spatial pyramid pooling are used to ensure accuracy and robustness. This method effectively reduces the network size and computational complexity and is a lightweight 6DoF pose estimation method based on monocular RGB images. Experiments on public datasets and self-built datasets show that the average ADD(-S) estimation accuracy and 2D projection index of the improved method are improved. For datasets with large changes in object scale, the estimation accuracy of the average ADD(-S) is greatly improved. |
| first_indexed | 2024-04-24T10:47:10Z |
| format | Article |
| id | doaj.art-df4a5f173bf94e73b4a3672fb54e7d5f |
| institution | Directory Open Access Journal |
| issn | 2079-9292 |
| language | English |
| last_indexed | 2024-04-24T10:47:10Z |
| publishDate | 2024-04-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Electronics |
| spelling | doaj.art-df4a5f173bf94e73b4a3672fb54e7d5f2024-04-12T13:17:23ZengMDPI AGElectronics2079-92922024-04-01137132110.3390/electronics13071321A Lightweight 6D Pose Estimation Network Based on Improved Atrous Spatial Pyramid PoolingFupan Wang0Xiaohang Tang1Yadong Wu2Yinfan Wang3Huarong Chen4Guijuan Wang5Jing Liao6School of Computer Science and Technology, Southwest University of Science and Technology, Mianyang 621010, ChinaSchool of Computer Science and Technology, Southwest University of Science and Technology, Mianyang 621010, ChinaSchool of Computer Science and Engineering, Sichuan University of Science & Engineering, Yibin 644007, ChinaSchool of Computer Science and Technology, Southwest University of Science and Technology, Mianyang 621010, ChinaSchool of Computer Science and Technology, Southwest University of Science and Technology, Mianyang 621010, ChinaSchool of Computer Science and Technology, Southwest University of Science and Technology, Mianyang 621010, ChinaSchool of Computer Science and Technology, Southwest University of Science and Technology, Mianyang 621010, ChinaIt is difficult for lightweight neural networks to produce accurate 6DoF pose estimation effects due to their accuracy being affected by scale changes. To solve this problem, we propose a method with good performance and robustness based on previous research. The enhanced PVNet-based method uses depth-wise convolution to build a lightweight network. In addition, coordinate attention and atrous spatial pyramid pooling are used to ensure accuracy and robustness. This method effectively reduces the network size and computational complexity and is a lightweight 6DoF pose estimation method based on monocular RGB images. Experiments on public datasets and self-built datasets show that the average ADD(-S) estimation accuracy and 2D projection index of the improved method are improved. For datasets with large changes in object scale, the estimation accuracy of the average ADD(-S) is greatly improved.https://www.mdpi.com/2079-9292/13/7/13216DoF pose estimationdepth-wise convolutioncoordinate attentionatrous spatial pyramid pooling |
| spellingShingle | Fupan Wang Xiaohang Tang Yadong Wu Yinfan Wang Huarong Chen Guijuan Wang Jing Liao A Lightweight 6D Pose Estimation Network Based on Improved Atrous Spatial Pyramid Pooling Electronics 6DoF pose estimation depth-wise convolution coordinate attention atrous spatial pyramid pooling |
| title | A Lightweight 6D Pose Estimation Network Based on Improved Atrous Spatial Pyramid Pooling |
| title_full | A Lightweight 6D Pose Estimation Network Based on Improved Atrous Spatial Pyramid Pooling |
| title_fullStr | A Lightweight 6D Pose Estimation Network Based on Improved Atrous Spatial Pyramid Pooling |
| title_full_unstemmed | A Lightweight 6D Pose Estimation Network Based on Improved Atrous Spatial Pyramid Pooling |
| title_short | A Lightweight 6D Pose Estimation Network Based on Improved Atrous Spatial Pyramid Pooling |
| title_sort | lightweight 6d pose estimation network based on improved atrous spatial pyramid pooling |
| topic | 6DoF pose estimation depth-wise convolution coordinate attention atrous spatial pyramid pooling |
| url | https://www.mdpi.com/2079-9292/13/7/1321 |
| work_keys_str_mv | AT fupanwang alightweight6dposeestimationnetworkbasedonimprovedatrousspatialpyramidpooling AT xiaohangtang alightweight6dposeestimationnetworkbasedonimprovedatrousspatialpyramidpooling AT yadongwu alightweight6dposeestimationnetworkbasedonimprovedatrousspatialpyramidpooling AT yinfanwang alightweight6dposeestimationnetworkbasedonimprovedatrousspatialpyramidpooling AT huarongchen alightweight6dposeestimationnetworkbasedonimprovedatrousspatialpyramidpooling AT guijuanwang alightweight6dposeestimationnetworkbasedonimprovedatrousspatialpyramidpooling AT jingliao alightweight6dposeestimationnetworkbasedonimprovedatrousspatialpyramidpooling AT fupanwang lightweight6dposeestimationnetworkbasedonimprovedatrousspatialpyramidpooling AT xiaohangtang lightweight6dposeestimationnetworkbasedonimprovedatrousspatialpyramidpooling AT yadongwu lightweight6dposeestimationnetworkbasedonimprovedatrousspatialpyramidpooling AT yinfanwang lightweight6dposeestimationnetworkbasedonimprovedatrousspatialpyramidpooling AT huarongchen lightweight6dposeestimationnetworkbasedonimprovedatrousspatialpyramidpooling AT guijuanwang lightweight6dposeestimationnetworkbasedonimprovedatrousspatialpyramidpooling AT jingliao lightweight6dposeestimationnetworkbasedonimprovedatrousspatialpyramidpooling |