DCVSMNet: Double Cost Volume Stereo Matching Network
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866909414546472960 |
|---|---|
| author | Tahmasebi, Mahmoud Huq, Saif Meehan, Kevin McAfee, Marion |
| author_facet | Tahmasebi, Mahmoud Huq, Saif Meehan, Kevin McAfee, Marion |
| contents | We introduce Double Cost Volume Stereo Matching Network(DCVSMNet) which is a novel architecture characterised by by two small upper (group-wise) and lower (norm correlation) cost volumes. Each cost volume is processed separately, and a coupling module is proposed to fuse the geometry information extracted from the upper and lower cost volumes. DCVSMNet is a fast stereo matching network with a 67 ms inference time and strong generalization ability which can produce competitive results compared to state-of-the-art methods. The results on several bench mark datasets show that DCVSMNet achieves better accuracy than methods such as CGI-Stereo and BGNet at the cost of greater inference time. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2402_16473 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | DCVSMNet: Double Cost Volume Stereo Matching Network Tahmasebi, Mahmoud Huq, Saif Meehan, Kevin McAfee, Marion Computer Vision and Pattern Recognition We introduce Double Cost Volume Stereo Matching Network(DCVSMNet) which is a novel architecture characterised by by two small upper (group-wise) and lower (norm correlation) cost volumes. Each cost volume is processed separately, and a coupling module is proposed to fuse the geometry information extracted from the upper and lower cost volumes. DCVSMNet is a fast stereo matching network with a 67 ms inference time and strong generalization ability which can produce competitive results compared to state-of-the-art methods. The results on several bench mark datasets show that DCVSMNet achieves better accuracy than methods such as CGI-Stereo and BGNet at the cost of greater inference time. |
| title | DCVSMNet: Double Cost Volume Stereo Matching Network |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2402.16473 |