Reproduce “AxisQ: Depth-Adaptive Channel--Frequency Quantization for Camouflaged 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
Chi Ruan, Tianqi Li, Xin He, Xue Geng, Ming-Ming Cheng, Yun Liu
AxisQ assigns activation bits by channel in shallow stages and spatial-frequency bands in deep stages of hierarchical camouflaged-object detectors, using task-sensitive costs under a fixed per-layer average bit budget. In CamoFormer-P on CAMO at W2A2, it raises weighted F-measure from LS-ViT’s 0.375 to 0.500. This supports depth-adaptive allocation in these settings, not transfer to other tasks or architectures.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.