6D Object Pose Estimation with Pairwise Compatible Geometric Features
This work addresses the problem of 6-DoF pose estimation under heavy occlusion. While previous work demonstrates reasonable results in unoccluded situations, robust and efficient pose estimation is still challenging in heavily occluded and low-texture scenarios which are ubiquitous in many applicati...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Published: |
IEEE
2021
|
| Online Access: | https://hdl.handle.net/1721.1/138123 |
| _version_ | 1826202595990962176 |
|---|---|
| author | Lin, Muyuan Murali, Varun Karaman, Sertac |
| author_facet | Lin, Muyuan Murali, Varun Karaman, Sertac |
| author_sort | Lin, Muyuan |
| collection | MIT |
| description | This work addresses the problem of 6-DoF pose estimation under heavy occlusion. While previous work demonstrates reasonable results in unoccluded situations, robust and efficient pose estimation is still challenging in heavily occluded and low-texture scenarios which are ubiquitous in many applications. To this end, we propose a novel end-to-end deep neural network model recovering object poses from depth measurements. The proposed model enforces pairwise consistency of 3D geometric features by applying spectral convolutions on a pairwise compatibility graph. We achieve comparable accuracy as the state-of-the-art graph matching solver while being much faster. Our approach outperforms state-of-the-art 6-DoF pose estimation methods on LineMOD and Occlusion LineMOD and runs in reasonable time (~5.9 Hz). We additionally verify this method on a synthetic dataset with large affine changes. |
| first_indexed | 2024-09-23T12:10:09Z |
| format | Article |
| id | mit-1721.1/138123 |
| institution | Massachusetts Institute of Technology |
| last_indexed | 2024-09-23T12:10:09Z |
| publishDate | 2021 |
| publisher | IEEE |
| record_format | dspace |
| spelling | mit-1721.1/1381232021-11-13T03:19:45Z 6D Object Pose Estimation with Pairwise Compatible Geometric Features Lin, Muyuan Murali, Varun Karaman, Sertac This work addresses the problem of 6-DoF pose estimation under heavy occlusion. While previous work demonstrates reasonable results in unoccluded situations, robust and efficient pose estimation is still challenging in heavily occluded and low-texture scenarios which are ubiquitous in many applications. To this end, we propose a novel end-to-end deep neural network model recovering object poses from depth measurements. The proposed model enforces pairwise consistency of 3D geometric features by applying spectral convolutions on a pairwise compatibility graph. We achieve comparable accuracy as the state-of-the-art graph matching solver while being much faster. Our approach outperforms state-of-the-art 6-DoF pose estimation methods on LineMOD and Occlusion LineMOD and runs in reasonable time (~5.9 Hz). We additionally verify this method on a synthetic dataset with large affine changes. 2021-11-12T15:28:30Z 2021-11-12T15:28:30Z 2021-05-30 Article http://purl.org/eprint/type/ConferencePaper https://hdl.handle.net/1721.1/138123 Lin, Muyuan, Murali, Varun and Karaman, Sertac. 2021. "6D Object Pose Estimation with Pairwise Compatible Geometric Features." 10.1109/icra48506.2021.9561404 Creative Commons Attribution-Noncommercial-Share Alike http://creativecommons.org/licenses/by-nc-sa/4.0/ application/pdf IEEE Muyuan Lin |
| spellingShingle | Lin, Muyuan Murali, Varun Karaman, Sertac 6D Object Pose Estimation with Pairwise Compatible Geometric Features |
| title | 6D Object Pose Estimation with Pairwise Compatible Geometric Features |
| title_full | 6D Object Pose Estimation with Pairwise Compatible Geometric Features |
| title_fullStr | 6D Object Pose Estimation with Pairwise Compatible Geometric Features |
| title_full_unstemmed | 6D Object Pose Estimation with Pairwise Compatible Geometric Features |
| title_short | 6D Object Pose Estimation with Pairwise Compatible Geometric Features |
| title_sort | 6d object pose estimation with pairwise compatible geometric features |
| url | https://hdl.handle.net/1721.1/138123 |
| work_keys_str_mv | AT linmuyuan 6dobjectposeestimationwithpairwisecompatiblegeometricfeatures AT muralivarun 6dobjectposeestimationwithpairwisecompatiblegeometricfeatures AT karamansertac 6dobjectposeestimationwithpairwisecompatiblegeometricfeatures |