AI Paper Atlas

Research map / Image generation / Image restoration

Breaking the Multi-Enhancement Bottleneck: Domain-Consistent Quality Enhancement for Compressed Images

Qunliang Xing, Ce Zheng, Mai Xu, Jing Yang, Shengxi Li

TPAMI 2025 · Image generation › Image restoration · approach: Low-light image enhancement · 0 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
adapt MPRNet with domain-consistent idempotent trainingMPRNet · DIV2K, BPGDI (PSNR, SSIM, TOPIQ, LPIPS): 59% PSNR, 46% SSIM, 52% TOPIQ, 74% LPIPS reductionImproved
adapt CBDNet with domain-consistent trainingCBDNet · DIV2K, BPGDI of PSNR (%): 3.25 -> 0.35Improved
adapt RBQE with domain-consistent trainingRBQE · DIV2K, BPGDI of PSNR (%): 3.50 -> 1.17Improved
adapt AR-CNN with domain-consistent trainingAR-CNN · DIV2K, BPGDI of PSNR (%): 1.85 -> 1.75Improved
adapt Real-ESRGAN with domain-consistent trainingReal-ESRGAN · DIV2K, BPGDI of PSNR (%): 6.88 -> 3.97Improved
adapt ESRGAN with domain-consistent trainingESRGAN · DIV2K, JPEGDI of PSNR (%): 7.04 -> 4.97Improved
adapt RBQE with domain-consistent trainingRBQE · DIV2K, JPEGDI of PSNR (%): 3.51 -> 0.12Improved
adapt methods with domain-consistent trainingall baselines · DIV2K and Flickr2K, single enhancementaverage percentage difference vs baseline: 0.197 PSNR, 0.145 SSIM, 1.445 LPIPS, 0.177 TOPIQNo clear effect
supervise multi-enhanced outputs with the raw image (straightforward solution) for CBDNet (ablation)CBDNet · DIV2K BPG QP37once-enhanced PSNR (dB): 31.66 -> 31.56Hurt
supervise multi-enhanced outputs with the raw image (straightforward solution) for CBDNet (ablation)CBDNet · DIV2K BPG QP37DI of PSNR after fifth enhancement (%): 3.19 -> 2.88Improved
supervise multi-enhanced outputs with the raw image (straightforward solution) for SRGAN (ablation)SRGAN · DIV2K BPG QP37DI of LPIPS after fifth enhancement (%): 79.22 -> 86.24Hurt
supervise multi-enhanced outputs with the raw image (straightforward solution) for SRGAN (ablation)SRGAN · DIV2K BPG QP37DI of PSNR after fifth enhancement (%): 8.76 -> 8.72No clear effect

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