File Download
There are no files associated with this item.
Links for fulltext
(May Require Subscription)
- Publisher Website: 10.1016/j.patcog.2021.108221
- Scopus: eid_2-s2.0-85112626949
- WOS: WOS:000697551500007
- Find via
Supplementary
- Citations:
- Appears in Collections:
Article: Learning scale awareness in keypoint extraction and description
Title | Learning scale awareness in keypoint extraction and description |
---|---|
Authors | |
Keywords | 3D reconstruction Image matching Keypoint description Keypoint detection Structure from motion |
Issue Date | 2022 |
Citation | Pattern Recognition, 2022, v. 121, article no. 108221 How to Cite? |
Abstract | To recover relative camera motion accurately and robustly, establishing a set of point-to-point correspondences in the pixel space is an essential yet challenging task in computer vision. Even though multi-scale design philosophy has been used with significant success in computer vision tasks, such as object detection and semantic segmentation, learning-based image matching has not been fully exploited. In this work, we explore a scale awareness learning approach in finding pixel-level correspondences based on the intuition that keypoints need to be extracted and described on an appropriate scale. With that insight, we propose a novel scale-aware network and then develop a new fusion scheme that derives high-consistency response maps and high-precision descriptions. We also revise the Second Order Similarity Regularization (SOSR) to make it more effective for the end-to-end image matching network, which leads to significant improvement in local feature descriptions. Experimental results run on multiple datasets demonstrate that our approach performs better than state-of-the-art methods under multiple criteria. |
Persistent Identifier | http://hdl.handle.net/10722/315203 |
ISSN | 2023 Impact Factor: 7.5 2023 SCImago Journal Rankings: 2.732 |
ISI Accession Number ID |
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Shen, Xuelun | - |
dc.contributor.author | Wang, Cheng | - |
dc.contributor.author | Li, Xin | - |
dc.contributor.author | Peng, Yifan | - |
dc.contributor.author | He, Zijian | - |
dc.contributor.author | Wen, Chenglu | - |
dc.contributor.author | Cheng, Ming | - |
dc.date.accessioned | 2022-08-05T10:18:02Z | - |
dc.date.available | 2022-08-05T10:18:02Z | - |
dc.date.issued | 2022 | - |
dc.identifier.citation | Pattern Recognition, 2022, v. 121, article no. 108221 | - |
dc.identifier.issn | 0031-3203 | - |
dc.identifier.uri | http://hdl.handle.net/10722/315203 | - |
dc.description.abstract | To recover relative camera motion accurately and robustly, establishing a set of point-to-point correspondences in the pixel space is an essential yet challenging task in computer vision. Even though multi-scale design philosophy has been used with significant success in computer vision tasks, such as object detection and semantic segmentation, learning-based image matching has not been fully exploited. In this work, we explore a scale awareness learning approach in finding pixel-level correspondences based on the intuition that keypoints need to be extracted and described on an appropriate scale. With that insight, we propose a novel scale-aware network and then develop a new fusion scheme that derives high-consistency response maps and high-precision descriptions. We also revise the Second Order Similarity Regularization (SOSR) to make it more effective for the end-to-end image matching network, which leads to significant improvement in local feature descriptions. Experimental results run on multiple datasets demonstrate that our approach performs better than state-of-the-art methods under multiple criteria. | - |
dc.language | eng | - |
dc.relation.ispartof | Pattern Recognition | - |
dc.subject | 3D reconstruction | - |
dc.subject | Image matching | - |
dc.subject | Keypoint description | - |
dc.subject | Keypoint detection | - |
dc.subject | Structure from motion | - |
dc.title | Learning scale awareness in keypoint extraction and description | - |
dc.type | Article | - |
dc.description.nature | link_to_subscribed_fulltext | - |
dc.identifier.doi | 10.1016/j.patcog.2021.108221 | - |
dc.identifier.scopus | eid_2-s2.0-85112626949 | - |
dc.identifier.volume | 121 | - |
dc.identifier.spage | article no. 108221 | - |
dc.identifier.epage | article no. 108221 | - |
dc.identifier.isi | WOS:000697551500007 | - |