Pose Estimation from RGB Images of Highly Symmetric Objects using a Novel Multi-Pose Loss and Differential Rendering
(2021) IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) 2021- Abstract
- We propose a novel multi-pose loss function to train a neural network for 6D pose estimation, using synthetic data and evaluating it on real images. Our loss is inspired by the VSD (Visible Surface Discrepancy) metric and relies on a differentiable renderer and CAD models. This novel multi-pose approach produces multiple weighted pose estimates to avoid getting stuck in local minima. Our method resolves pose ambiguities without using predefined symmetries. It is trained only on synthetic data. We test on real-world RGB images from the T-LESS dataset, containing highly symmetric objects common in industrial settings. We show that our solution can be used to replace the codebook in a state-of-the-art approach. So far, the codebook approach... (More)
- We propose a novel multi-pose loss function to train a neural network for 6D pose estimation, using synthetic data and evaluating it on real images. Our loss is inspired by the VSD (Visible Surface Discrepancy) metric and relies on a differentiable renderer and CAD models. This novel multi-pose approach produces multiple weighted pose estimates to avoid getting stuck in local minima. Our method resolves pose ambiguities without using predefined symmetries. It is trained only on synthetic data. We test on real-world RGB images from the T-LESS dataset, containing highly symmetric objects common in industrial settings. We show that our solution can be used to replace the codebook in a state-of-the-art approach. So far, the codebook approach has had the shortest inference time in the field. Our approach reduces inference time further while a) avoiding discretization, b) requiring a much smaller memory footprint and c) improving pose recall. (Less)
Please use this url to cite or link to this publication:
https://lup.lub.lu.se/record/d3b021c6-d263-4bf6-89f5-3c7cb1637083
- author
- Hein Bengtson, Stefan LU ; Åström, Hampus LU ; Moeslund, Thomas B. ; Topp, Elin Anna LU and Krueger, Volker LU
- organization
- publishing date
- 2021-09
- type
- Chapter in Book/Report/Conference proceeding
- publication status
- published
- subject
- keywords
- Training, Measurement, Solid modeling, image resolution, Pose estimation, neural networks, memory management
- host publication
- IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)
- edition
- 2021
- publisher
- IEEE - Institute of Electrical and Electronics Engineers Inc.
- conference name
- IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) 2021
- conference location
- Prague, Czech Republic
- conference dates
- 2021-09-27 - 2021-10-01
- external identifiers
-
- scopus:85124360826
- ISBN
- 978-1-6654-1715-0
- 978-1-6654-1714-3
- DOI
- 10.1109/IROS51168.2021.9636839
- project
- RobotLab LTH
- language
- English
- LU publication?
- yes
- id
- d3b021c6-d263-4bf6-89f5-3c7cb1637083
- date added to LUP
- 2022-07-14 23:44:08
- date last changed
- 2024-08-22 22:19:22
@inproceedings{d3b021c6-d263-4bf6-89f5-3c7cb1637083, abstract = {{We propose a novel multi-pose loss function to train a neural network for 6D pose estimation, using synthetic data and evaluating it on real images. Our loss is inspired by the VSD (Visible Surface Discrepancy) metric and relies on a differentiable renderer and CAD models. This novel multi-pose approach produces multiple weighted pose estimates to avoid getting stuck in local minima. Our method resolves pose ambiguities without using predefined symmetries. It is trained only on synthetic data. We test on real-world RGB images from the T-LESS dataset, containing highly symmetric objects common in industrial settings. We show that our solution can be used to replace the codebook in a state-of-the-art approach. So far, the codebook approach has had the shortest inference time in the field. Our approach reduces inference time further while a) avoiding discretization, b) requiring a much smaller memory footprint and c) improving pose recall.}}, author = {{Hein Bengtson, Stefan and Åström, Hampus and Moeslund, Thomas B. and Topp, Elin Anna and Krueger, Volker}}, booktitle = {{IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)}}, isbn = {{978-1-6654-1715-0}}, keywords = {{Training; Measurement; Solid modeling; image resolution; Pose estimation; neural networks; memory management}}, language = {{eng}}, publisher = {{IEEE - Institute of Electrical and Electronics Engineers Inc.}}, title = {{Pose Estimation from RGB Images of Highly Symmetric Objects using a Novel Multi-Pose Loss and Differential Rendering}}, url = {{http://dx.doi.org/10.1109/IROS51168.2021.9636839}}, doi = {{10.1109/IROS51168.2021.9636839}}, year = {{2021}}, }