AI Paper Atlas

Research map / Image generation / GANs

Deep Generative Adversarial Compression Artifact Removal

Leonardo Galteri, Lorenzo Seidenari, Marco Bertini, Alberto Del Bimbo

ICCV 2017 · Image generation › GANs · approach: GAN-based image manipulation · 224 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 the GAN with perceptual loss to the residual generator for object detectionResidual generator · PASCAL VOC2007 detection (Faster R-CNN), JPEG QF=20 restoredmAP: 0.574 -> 0.623Improved
train with SSIM loss instead of MSEResidual generator · LIVE1, JPEG QF=10SSIM: 0.834 -> 0.840Improved
train with SSIM loss instead of MSEResidual generator · LIVE1, JPEG QF=10PSNR (dB): 29.45 -> 28.94Hurt
add adversarial training (GAN) instead of MSEResidual generator · LIVE1, JPEG QF=10PSNR (dB): 29.45 -> 27.29Hurt
use GAN instead of SSIM loss for perceived qualityResidual generator · BSD500 subjective studyMOS: 49.51 -> 68.32Improved
use perceptual loss l_P instead of simple L1 loss in GAN (ablation)Our GAN · PASCAL VOC2007 detection (Faster R-CNN), JPEG QF=20 restoredmAP: 0.598 -> 0.623Improved
add color (RGB) instead of luminance only in GAN (ablation)Our GAN · PASCAL VOC2007 detection (Faster R-CNN), JPEG QF=20 restoredmAP: 0.598 -> 0.623Improved
use sub-patch discriminator instead of full patch discriminator (ablation)Our GAN · PASCAL VOC2007 detection (Faster R-CNN), JPEG QF=20 restoredmAP: 0.605 -> 0.623Improved
restore compressed images with MSE/SSIM generator before detectionFaster R-CNN · PASCAL VOC2007, classes cat, cow, dog, sheepper-class APHurt
apply GAN restoration to WebP-compressed imagesFaster R-CNN · PASCAL VOC2007, WebPmAP: +4%Improved
apply GAN restoration to JPEG2000 imagesFaster R-CNN · PASCAL VOC2007, JPEG2000mAP: +2.9%Improved
apply GAN restoration to BPG imagesFaster R-CNN · PASCAL VOC2007, BPGmAP: +2.1%Improved

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