AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

CoopDiff: A Diffusion-Guided Approach for Cooperation under Corruptions

Gong Chen, Chaokun Zhang, Pengcheng Lv

CVPR 2026 · Image generation › Fast diffusion sampling · approach: Training-free diffusion acceleration · 3 citations (OpenAlex)

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 GCM-driven diffusion to baseline (ablation)CoopDiff baseline · OPV2VnAP@0.5 / AP@0.7: 0.8548 -> 0.8681 (AP@0.5)Improved
add all modules (GCM, Coop, Teacher, EGCA) (ablation)baseline · OPV2VnAP@0.5: 0.8548 -> 0.9053Improved
add QoI weighting to naive early-fusion teacher (ablation)teacher · OPV2VnAP@0.7: 0.7717 -> 0.8258Improved
add semantic guidance mask to teacher (ablation)teacher with QoI · OPV2VnAP@0.7: 0.8258 -> 0.8325Improved
add GCM to teacher (ablation)teacher with QoI+SGM · OPV2VnAP@0.7: 0.8325 -> 0.8357Improved
increase diffusion steps from 1 to 2 (ablation)CoopDiff · OPV2VnAP@0.7: 0.8152 -> 0.8326Improved
increase diffusion steps from 10 to 20 (ablation)CoopDiff · OPV2VnAP@0.7 and FPS: 0.8357 -> 0.8362; FPS 5.31 -> 3.22No clear effect
reduce selection ratio from 30% to 5% (ablation)cooperative branch · DAIR-V2XnAP@0.7: 0.6644 -> 0.6344Hurt
raise selection ratio from 30% to 50% (ablation)cooperative branch · DAIR-V2XnAP@0.7: 0.6644 -> 0.6642No clear effect
use CoopDiff instead of DSRCcooperative perception · OPV2VnmRCE: 15.77 -> 12.94Improved
use CoopDiff instead of DSRCcooperative perception · Culver City unseen domainAP@0.7 drop: DSRC 11.94% vs ours 09.50% drop; V2VNet 09.05% dropMixed

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