AI Paper Atlas

Research map / Image generation / Image restoration

Bitstream-Corrupted JPEG Images Are Restorable: Two-Stage Compensation and Alignment Framework for Image Restoration

Wenyang Liu, Yi Wang, Kim-Hui Yap, Lap-Pui Chau

CVPR 2023 · Image generation › Image restoration · approach: Efficient restoration architectures · 13 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 SCA before GCA, including segment compensation and block alignmentJPEG restoration framework · Cityscapes 512×256PSNR: 12.41 -> 14.34Improved
Use GCA instead of EPDN after SCASCA restoration output · Cityscapes 512×256PSNR: 33.51 -> 42.05Improved
Use GCA instead of EPDN after SCASCA restoration output · Cityscapes 1024×512PSNR: 33.25 -> 40.09Improved
Use GCA instead of EPDN after SCASCA restoration output · Cityscapes 2048×1024PSNR: 33.40 -> 38.92Improved
Add SCA segment detection and normalization (ablation)SCA + GCA · Cityscapes 512×256PSNR: 33.88 -> 42.05Improved
Add SCA segment detection and normalization (ablation)SCA + GCA · Cityscapes 1024×512PSNR: 33.52 -> 40.09Improved
Add block alignment to SCA (ablation)SCA + GCA · Cityscapes 512×256PSNR: 36.50 -> 42.05Improved
Add block alignment to SCA (ablation)SCA + GCA · Cityscapes 1024×512PSNR: 35.01 -> 40.09Improved
Use Laplacian fusion instead of EPDN's pooling-enhanced module (ablation)GCA · Cityscapes 512×256PSNR: 36.65 -> 42.05Improved
Use Laplacian fusion instead of EPDN's pooling-enhanced module (ablation)GCA · Cityscapes 1024×512PSNR: 36.16 -> 40.09Improved
Add edge and Charbonnier losses instead of using original ℓ2 fidelity loss (ablation)GCA training · Cityscapes 512×256PSNR: 41.27 -> 42.05Improved
Add edge and Charbonnier losses instead of using original ℓ2 fidelity loss (ablation)GCA training · Cityscapes 1024×512PSNR: 39.53 -> 40.09Improved

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