AI Paper Atlas

Research map / Image generation / Image restoration

Uncover Treasures in DCT: Advancing JPEG Quality Enhancement by Exploiting Latent Correlations

Jing Yang, Qunliang Xing, Mai Xu, Minglang Qiao

ICCV 2025 · Image generation › Image restoration · approach: Perceptual diffusion image compression · 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 block-based convolution EMBC modules instead of ResBlocks (ablation)AJQE w/ FBCNN · Color JPEG, QF10PSNR (dB): 27.29 -> 27.69Improved
add the point-based correlation module EMPC instead of a ResBlock (ablation)AJQE w/ FBCNN · Color JPEG, QF10PSNR (dB): 27.59 -> 27.69Improved
use EMBC and EMPC together instead of ResBlocks for both (ablation)AJQE w/ FBCNN · Color JPEG, QF10PSNR (dB): 26.92 -> 27.69Improved
use block-based convolution instead of vanilla convolution in EMBC (ablation)AJQE w/ FBCNN · Color JPEGPSNR (dB): 0.14 dB decrease in PSNR for BConv-VConvImproved
use UMEP instead of spatial nearest-neighbor upsampling of chroma DCT (ablation)AJQE w/ FBCNN · Color JPEGPSNR (dB): average decrease in PSNR of 0.16 dBImproved
use UMEP instead of transposed-convolution upsampling of chroma DCT (ablation)AJQE w/ FBCNN · Color JPEGPSNR (dB): average decrease in PSNR of 0.10 dBImproved
use UMEP instead of spatial nearest-neighbor upsampling at QF40 (ablation)AJQE w/ FBCNN · Color JPEG, QF40PSNR (dB): 32.24 -> 32.41Improved
adapt pixel-domain model to the DCT domain with AJQEpixel-domain models · Color JPEGPSNR (dB): 0.35 dB improvement in PSNRImproved
adapt pixel-domain model to the DCT domain with AJQEpixel-domain models · Grayscale JPEGPSNR (dB): 0.39 dB improvement in PSNRImproved
AJQE versus existing DCT-domain methodsQGAC and JDEC · Color JPEGPSNR (dB): 0.41 dB improvementImproved
AJQE versus pixel-domain models on efficiencypixel-domain models · BSDS500enhancement throughput: 60.5% higher enhancement throughputImproved
AJQE versus pixel-domain models on computepixel-domain models · BSDS500FLOPs: 45.9% fewer FLOPsImproved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 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