AI Paper Atlas

Research map / Video / Video segmentation and tracking

Separation for Better Integration: Disentangling Edge and Motion in Event-based Deblurring

Yufei Zhu, Hao Chen, Yongjian Deng, Wei You

ICCV 2025 · Video › Video segmentation and tracking · approach: Event-based video reconstruction · 1 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 plain cross-attention event fusion (ablation)two-stage U-Net baseline · EVRBPSNR: 28.65 -> 30.71Improved
add edge-aware sharpening module (ablation)baseline · EVRBPSNR: 28.65 -> 31.13Improved
add edge-aware sharpening instead of plain event attention (ablation)baseline with events · EVRBPSNR: 30.71 -> 31.13Improved
add deformable conv with offsets from image features (ablation)baseline · EVRBSSIM: 28.65 -> 28.80No clear effect
add deformable conv with event-feature offsets (ablation)baseline · EVRBPSNR: 28.65 -> 30.51Improved
add motion-driven deformable conv using motion cues (ablation)baseline · EVRBPSNR: 28.65 -> 31.63Improved
use motion cues instead of full event features for DCN offsets (ablation)baseline with event DCN · EVRBPSNR: 30.51 -> 31.63Improved
combine edge-aware sharpening with motion-driven deblurring (ablation)B+MotionDCN · EVRBPSNR: 31.63 -> 32.05Improved
combine motion-driven deblurring with edge-aware sharpening (ablation)B+EdgeAtt · EVRBPSNR: 31.13 -> 32.05Improved
compare single-frame model with multi-frame CMTA-5Ours · GoProPSNR: 36.55 -> 36.70Improved
compare single-frame model with CMTA-7Ours · GoProPSNR: 36.78 -> 36.70Hurt
compare with FEVD multi-frame on REVDOurs · REVDPSNR: 32.99 -> 32.14Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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