JFRDet aligns infrared and visible features with an explicit affine transformation and illumination-guided fusion, reaching 69.7% mAP50 on the new DVMA benchmark.
Beyond bounding-box: Convex- hull feature adaptation for oriented and densely packed ob- ject detection
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Bridging Severe Cross-Modal Misalignment: End-to-End Visible-Infrared Object Detection via Explicit Feature-Domain Affine Registration
JFRDet aligns infrared and visible features with an explicit affine transformation and illumination-guided fusion, reaching 69.7% mAP50 on the new DVMA benchmark.