.png){kind=icon}
.png){kind=icon}
Return
Parallel consensus transformer for local feature matching
.png){kind=icon}
PRE
AI
DOI：10.1016/j.patcog.2025.112905.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• A local consensus module is introduced to reduce mismatches via local cues. • A parallel consensus attention layer jointly captures global context and local structure. • A lightweight multi-scale descriptor is built using frozen SuperPoint features. • PCMatcher outperforms SOTA on various benchmarks for feature matching tasks.
Journal
IF:
7.6
Papers:
1.3W
Citations:
4.5W