AI Paper Atlas

Research map / Architectures and efficiency / CNN design

JPEG Artifacts Reduction via Deep Convolutional Sparse Coding

Xueyang Fu, Zheng-Jun Zha, Feng Wu, Xinghao Ding, John Paisley

ICCV 2019 · Architectures and efficiency › CNN design · approach: CNN filter pruning · 128 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 dilated convolutions for multi-scale features (ablation)DCSC · LIVE1 (JPEG quality 10)PSNR-B: 29.08 -> 29.17Improved
add dilated convolutions for multi-scale features (ablation)DCSC · LIVE1 (JPEG quality 20)PSNR-B: 31.39 -> 31.47Improved
add dilated convolutions for multi-scale features (ablation)DCSC · LIVE1 (JPEG quality 30)PSNR-B: 32.78 -> 32.81Improved
increase iterations from R = 20 to R = 40 (64 filters) (ablation)DCSC · LIVE1 (JPEG quality 10)PSNR: 28.46 -> 29.17Improved
increase iterations from R = 40 to R = 60 (64 filters) (ablation)DCSC · LIVE1 (JPEG quality 10)PSNR: 29.17 -> 29.21Improved
increase filters from 32 to 128 (R = 40) (ablation)DCSC · LIVE1 (JPEG quality 10)PSNR: 29.14 -> 29.19Improved
increase filters from 32 to 128 (R = 20) (ablation)DCSC · LIVE1 (JPEG quality 10)PSNR: 28.34 -> 28.67Improved
use MAE loss instead of MSE loss (ablation)DCSC · LIVE1 (JPEG quality 10)PSNR: 29.13 -> 29.17Improved
use MAE loss instead of MSE loss (ablation)DCSC · LIVE1 (JPEG quality 20)PSNR-B: 31.33 -> 31.47Improved
use MAE loss instead of MSE loss (ablation)DCSC · LIVE1 (JPEG quality 30)PSNR: 32.70 -> 32.83Improved
use DCSC instead of DnCNN (86.49% fewer parameters)JPEG artifact reduction network · BSD500 (JPEG quality 30)PSNR-B: 31.43 -> 32.19Improved
use DCSC instead of DnCNNJPEG artifact reduction network · Classic5 (JPEG quality 10)PSNR: 29.40 -> 29.25No 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