SSL-Net: Point-Cloud Generation Network With Self-Supervised Learning
Inferring the three-dimensional structure of objects from monocular images has far-reaching applications in the field of 3D perception. In this paper, we propose a self-supervised network (SSL-Net) to generate 3D point clouds from a single RGB image, unlike the existing work which requires multiple...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2019-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/8740985/ |
| _version_ | 1818922125433503744 |
|---|---|
| author | Ran Sun Yongbin Gao Zhijun Fang Anjie Wang Cengsi Zhong |
| author_facet | Ran Sun Yongbin Gao Zhijun Fang Anjie Wang Cengsi Zhong |
| author_sort | Ran Sun |
| collection | DOAJ |
| description | Inferring the three-dimensional structure of objects from monocular images has far-reaching applications in the field of 3D perception. In this paper, we propose a self-supervised network (SSL-Net) to generate 3D point clouds from a single RGB image, unlike the existing work which requires multiple views of the same object to recover the full 3D geometry. To provide the extra self-supervisory signal, the generated 3D model is simultaneously rendered into an image and compared with the input image. In addition, a pose estimation network is integrated into the 3D point cloud generation network to eliminate the pose ambiguity of the input image, and the estimated pose is also used for rendering the 2D image with the same pose as input image from 3D point clouds. The extensive experiments on both real and synthetic datasets show that our method not only qualitatively generates point clouds with more details but also quantitatively outperforms the state-of-the-art in accuracy. |
| first_indexed | 2024-12-20T01:48:34Z |
| format | Article |
| id | doaj.art-1c86fd99ba3c483fbc1c2f9db28e1585 |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-12-20T01:48:34Z |
| publishDate | 2019-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-1c86fd99ba3c483fbc1c2f9db28e15852022-12-21T19:57:41ZengIEEEIEEE Access2169-35362019-01-017822068221710.1109/ACCESS.2019.29238428740985SSL-Net: Point-Cloud Generation Network With Self-Supervised LearningRan Sun0https://orcid.org/0000-0001-8563-5678Yongbin Gao1Zhijun Fang2Anjie Wang3https://orcid.org/0000-0002-9486-1009Cengsi Zhong4School of Electronic and Electrical Engineering, Shanghai University of Engineering Science, Shanghai, ChinaSchool of Electronic and Electrical Engineering, Shanghai University of Engineering Science, Shanghai, ChinaSchool of Electronic and Electrical Engineering, Shanghai University of Engineering Science, Shanghai, ChinaSchool of Electronic and Electrical Engineering, Shanghai University of Engineering Science, Shanghai, ChinaSchool of Electronic and Electrical Engineering, Shanghai University of Engineering Science, Shanghai, ChinaInferring the three-dimensional structure of objects from monocular images has far-reaching applications in the field of 3D perception. In this paper, we propose a self-supervised network (SSL-Net) to generate 3D point clouds from a single RGB image, unlike the existing work which requires multiple views of the same object to recover the full 3D geometry. To provide the extra self-supervisory signal, the generated 3D model is simultaneously rendered into an image and compared with the input image. In addition, a pose estimation network is integrated into the 3D point cloud generation network to eliminate the pose ambiguity of the input image, and the estimated pose is also used for rendering the 2D image with the same pose as input image from 3D point clouds. The extensive experiments on both real and synthetic datasets show that our method not only qualitatively generates point clouds with more details but also quantitatively outperforms the state-of-the-art in accuracy.https://ieeexplore.ieee.org/document/8740985/3D reconstructionpoint-cloudself-supervised learning3D shape completionsingle view reconstruction |
| spellingShingle | Ran Sun Yongbin Gao Zhijun Fang Anjie Wang Cengsi Zhong SSL-Net: Point-Cloud Generation Network With Self-Supervised Learning IEEE Access 3D reconstruction point-cloud self-supervised learning 3D shape completion single view reconstruction |
| title | SSL-Net: Point-Cloud Generation Network With Self-Supervised Learning |
| title_full | SSL-Net: Point-Cloud Generation Network With Self-Supervised Learning |
| title_fullStr | SSL-Net: Point-Cloud Generation Network With Self-Supervised Learning |
| title_full_unstemmed | SSL-Net: Point-Cloud Generation Network With Self-Supervised Learning |
| title_short | SSL-Net: Point-Cloud Generation Network With Self-Supervised Learning |
| title_sort | ssl net point cloud generation network with self supervised learning |
| topic | 3D reconstruction point-cloud self-supervised learning 3D shape completion single view reconstruction |
| url | https://ieeexplore.ieee.org/document/8740985/ |
| work_keys_str_mv | AT ransun sslnetpointcloudgenerationnetworkwithselfsupervisedlearning AT yongbingao sslnetpointcloudgenerationnetworkwithselfsupervisedlearning AT zhijunfang sslnetpointcloudgenerationnetworkwithselfsupervisedlearning AT anjiewang sslnetpointcloudgenerationnetworkwithselfsupervisedlearning AT cengsizhong sslnetpointcloudgenerationnetworkwithselfsupervisedlearning |