AI Paper Atlas

Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning

More Than Meets the Eye: A Unified Image Fusion Framework via Semantic-Pixel Entropy Trade-off for Zero-Shot Generalization

Xiaowen Liu, Jing Li, Hongtao Huo, Haozhe Cao, Renhua Wang, Xu Dong

CVPR 2026 · Label-efficient and robust learning › Zero-/few-shot and meta-learning · approach: Transductive and semantic-dictionary zero-shot

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add reconstruction path to perception-path-only training (ablation)DECC · M3FD (YOLOv8 detection)mAP@[0.5:0.95]: 34.28 -> 36.34Improved
add perception path to reconstruction-path-only training (ablation)DECC · M3FD (YOLOv8 detection)mAP@[0.5:0.95]: 28.93 -> 36.34Improved
add perception path to reconstruction-path-only training (ablation)DECC · M3FD fusion qualityPSNR: 11.39 -> 16.48Improved
add reconstruction path to perception-path-only training (ablation)DECC · M3FD fusion qualitySD: 39.62 -> 40.97Improved
add KL trade-off loss (ablation)DECC · M3FD (YOLOv8 detection)mAP@[0.5:0.95]: 34.61 -> 36.34Improved
add KL trade-off loss (ablation)DECC · M3FD (YOLOv11 detection)mAP@[0.5:0.95]: 36.01 -> 36.85Improved
add grouped frequency dynamic convolution (GFDM) instead of conventional convolution (ablation)DECC · M3FD (YOLOv8 detection)mAP@[0.5:0.95]: 32.76 -> 36.34Improved
add GFDM instead of conventional convolution (ablation)DECC · M3FD fusion qualityEN: 6.49 -> 7.01Improved
reconstruction path only versus full dual-path on entropy (ablation)DECC · M3FD fusion qualityEN: only RP 7.07 vs DECC 7.01Mixed
add perception path to reconstruction-only training (ablation)DECC · M3FD fusion qualityMI: only PP 4.25, DECC 3.89, only RP 3.20Mixed
compare DECC to best baseline for detection on fused images (YOLOv8)DECC · M3FD (YOLOv8 detection)mAP@0.5: 59.36 -> 60.17Improved
compare DECC to best baseline for segmentation on fused images (Mask2Former)DECC · FMB (Mask2Former)mIoU: 55.28 -> 55.39Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy