Efficient pose estimation from single RGB-D image via Hough forest with auto-context
We propose a high efficient learning approach to estimating 6D (Degree of Freedom) pose of the textured or texture-less objects for grasping purposes in a cluttered environment where the objects might be partially occluded. The method comprises three main steps. Given a single RGB-D image, we first...
| Main Authors: | , , , , , |
|---|---|
| Other Authors: | |
| Format: | Conference Paper |
| Language: | English |
| Published: |
2020
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/143044 |
| _version_ | 1826129687098687488 |
|---|---|
| author | Dong, Huixu Prasad, Dilip Kumar Yuan, Qilong Zhou, Jiadong Asadi, Ehsan Chen, I-Ming |
| author2 | School of Computer Science and Engineering |
| author_facet | School of Computer Science and Engineering Dong, Huixu Prasad, Dilip Kumar Yuan, Qilong Zhou, Jiadong Asadi, Ehsan Chen, I-Ming |
| author_sort | Dong, Huixu |
| collection | NTU |
| description | We propose a high efficient learning approach to estimating 6D (Degree of Freedom) pose of the textured or texture-less objects for grasping purposes in a cluttered environment where the objects might be partially occluded. The method comprises three main steps. Given a single RGB-D image, we first deploy appropriate features and the random forest to deduce the object class probability and cast votes for the 6D pose in Hough space by joint regression and classification framework, adopting reservoir sampling and summarizing the pose distribution by clustering. Next, we integrate the auto-context into cascaded Hough forests to improve the efficiency of learning. Extensive experiments on various public datasets and robotic grasps indicate that our method presents some improvements over the state-of-art and reveals the capability for estimating poses in practical applications efficiently. |
| first_indexed | 2024-10-01T07:44:18Z |
| format | Conference Paper |
| id | ntu-10356/143044 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T07:44:18Z |
| publishDate | 2020 |
| record_format | dspace |
| spelling | ntu-10356/1430442023-03-04T17:08:05Z Efficient pose estimation from single RGB-D image via Hough forest with auto-context Dong, Huixu Prasad, Dilip Kumar Yuan, Qilong Zhou, Jiadong Asadi, Ehsan Chen, I-Ming School of Computer Science and Engineering School of Materials Science and Engineering School of Mechanical and Aerospace Engineering 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) Engineering::Mechanical engineering Training Forestry We propose a high efficient learning approach to estimating 6D (Degree of Freedom) pose of the textured or texture-less objects for grasping purposes in a cluttered environment where the objects might be partially occluded. The method comprises three main steps. Given a single RGB-D image, we first deploy appropriate features and the random forest to deduce the object class probability and cast votes for the 6D pose in Hough space by joint regression and classification framework, adopting reservoir sampling and summarizing the pose distribution by clustering. Next, we integrate the auto-context into cascaded Hough forests to improve the efficiency of learning. Extensive experiments on various public datasets and robotic grasps indicate that our method presents some improvements over the state-of-art and reveals the capability for estimating poses in practical applications efficiently. Accepted version 2020-07-23T02:05:28Z 2020-07-23T02:05:28Z 2019 Conference Paper Dong, H., Prasad, D. K., Yuan, Q., Zhou, J., Asadi, E., & Chen, I.-M. (2018). Efficient pose estimation from single RGB-D image via Hough forest with auto-context. Proceedings of 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 7201-7206. doi:10.1109/iros.2018.8594064 978-1-5386-8095-7 https://hdl.handle.net/10356/143044 10.1109/iros.2018.8594064 2-s2.0-85062992026 7201 7206 en © 2018 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. The published version is available at: https://doi.org/10.1109/iros.2018.8594064 application/pdf |
| spellingShingle | Engineering::Mechanical engineering Training Forestry Dong, Huixu Prasad, Dilip Kumar Yuan, Qilong Zhou, Jiadong Asadi, Ehsan Chen, I-Ming Efficient pose estimation from single RGB-D image via Hough forest with auto-context |
| title | Efficient pose estimation from single RGB-D image via Hough forest with auto-context |
| title_full | Efficient pose estimation from single RGB-D image via Hough forest with auto-context |
| title_fullStr | Efficient pose estimation from single RGB-D image via Hough forest with auto-context |
| title_full_unstemmed | Efficient pose estimation from single RGB-D image via Hough forest with auto-context |
| title_short | Efficient pose estimation from single RGB-D image via Hough forest with auto-context |
| title_sort | efficient pose estimation from single rgb d image via hough forest with auto context |
| topic | Engineering::Mechanical engineering Training Forestry |
| url | https://hdl.handle.net/10356/143044 |
| work_keys_str_mv | AT donghuixu efficientposeestimationfromsinglergbdimageviahoughforestwithautocontext AT prasaddilipkumar efficientposeestimationfromsinglergbdimageviahoughforestwithautocontext AT yuanqilong efficientposeestimationfromsinglergbdimageviahoughforestwithautocontext AT zhoujiadong efficientposeestimationfromsinglergbdimageviahoughforestwithautocontext AT asadiehsan efficientposeestimationfromsinglergbdimageviahoughforestwithautocontext AT cheniming efficientposeestimationfromsinglergbdimageviahoughforestwithautocontext |