| _version_ |
1826214709831925760
|
| collection |
MIT
|
| description |
© 2019 Neural information processing systems foundation. All rights reserved. We present a simple, flexible, and general framework titled Partial Registration Network (PRNet), for partial-to-partial point cloud registration. Inspired by recently-proposed learning-based methods for registration, we use deep networks to tackle non-convexity of the alignment and partial correspondence problems. While previous learning-based methods assume the entire shape is visible, PRNet is suitable for partial-to-partial registration, outperforming PointNetLK, DCP, and non-learning methods on synthetic data. PRNet is self-supervised, jointly learning an appropriate geometric representation, a keypoint detector that finds points in common between partial views, and keypoint-to-keypoint correspondences. We show PRNet predicts keypoints and correspondences consistently across views and objects. Furthermore, the learned representation is transferable to classification.
|
| first_indexed |
2024-09-23T16:09:48Z
|
| format |
Article
|
| id |
mit-1721.1/137356
|
| institution |
Massachusetts Institute of Technology
|
| language |
English
|
| last_indexed |
2024-09-23T16:09:48Z
|
| publishDate |
2021
|
| record_format |
dspace
|
| spelling |
mit-1721.1/1373562022-04-01T17:47:12Z PRNet: Self-supervised learning for partial-to-partial registration © 2019 Neural information processing systems foundation. All rights reserved. We present a simple, flexible, and general framework titled Partial Registration Network (PRNet), for partial-to-partial point cloud registration. Inspired by recently-proposed learning-based methods for registration, we use deep networks to tackle non-convexity of the alignment and partial correspondence problems. While previous learning-based methods assume the entire shape is visible, PRNet is suitable for partial-to-partial registration, outperforming PointNetLK, DCP, and non-learning methods on synthetic data. PRNet is self-supervised, jointly learning an appropriate geometric representation, a keypoint detector that finds points in common between partial views, and keypoint-to-keypoint correspondences. We show PRNet predicts keypoints and correspondences consistently across views and objects. Furthermore, the learned representation is transferable to classification. 2021-11-04T16:18:24Z 2021-11-04T16:18:24Z 2019-12 2019-12 2021-03-26T14:01:18Z Article http://purl.org/eprint/type/ConferencePaper https://hdl.handle.net/1721.1/137356 2019. "PRNet: Self-supervised learning for partial-to-partial registration." Advances in Neural Information Processing Systems, 32. en Advances in Neural Information Processing Systems Article is made available in accordance with the publisher's policy and may be subject to US copyright law. Please refer to the publisher's site for terms of use. application/pdf Neural Information Processing Systems (NIPS)
|
| spellingShingle |
PRNet: Self-supervised learning for partial-to-partial registration
|
| title |
PRNet: Self-supervised learning for partial-to-partial registration
|
| title_full |
PRNet: Self-supervised learning for partial-to-partial registration
|
| title_fullStr |
PRNet: Self-supervised learning for partial-to-partial registration
|
| title_full_unstemmed |
PRNet: Self-supervised learning for partial-to-partial registration
|
| title_short |
PRNet: Self-supervised learning for partial-to-partial registration
|
| title_sort |
prnet self supervised learning for partial to partial registration
|
| url |
https://hdl.handle.net/1721.1/137356
|