Reproduce “ILLRAW: Illumination-Aware RAW Object Detection” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Xinzhe Wang, Xiao Liang, Xiaoli Zhang, Kang Ma
ILLRAW adapts RAW camera data directly for object detection, replacing a detector’s initial embedding layer while retaining its downstream network. Its lightweight adapter estimates a spatial illumination mask and forms channel-sensitive features that are fused before detection. On AODRaw, Cascade-RCNN achieves 23.3 mAP, versus 21.1 for RAOD; this result is specific to that benchmark and detector.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.