AI Paper Atlas

Research map / Image generation / Image restoration

Unsupervised Variational Translator for Bridging Image Restoration and High-Level Vision Tasks

Jiawei Wu, Zhi Jin

ECCV 2024 · Image generation › Image restoration · approach: Priors for restoration networks · 5 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 maximum likelihood (L_mle) term to the cycle loss (ablation)VaT · RTTS dehazing detectionmAP %: 11.45 -> 42.77Improved
add cycle-consistency loss to L_mle (ablation)VaT · RTTS dehazing detectionmAP %: 41.97 -> 42.77Improved
add gated fusion module (ablation)VaT · RTTS dehazing detectionmAP %: 42.77 -> 47.49Improved
add gated fusion module (ablation)VaT · Low-light CIFAR-10 CLIPaccuracy %: 35.10 -> 41.62Improved
add Mixup augmentation (ablation)VaT · RTTS dehazing detectionmAP %: 47.49 -> 49.12Improved
use uncertainty-guided UMix instead of plain Mixup (ablation)VaT · Low-light CIFAR-10 CLIPaccuracy %: 43.19 -> 44.27Improved
use a stronger restoration network (DCP, AODNet, GCANet by PSNR) (ablation)VaT · VOCh-testdetection mAPImproved
increase translator size (0.21M to 6.10M parameters) (ablation)VaT · RTTSdetection mAPImproved
apply VaT over unsupervised translation methods (CycleGAN, CUT, MSPC, SRC, SPR)AODNet + YOLOv3 · RTTSmAP: 33.51 -> 49.48Improved
VaT versus supervised AD-YOLO on real hazy dataVaT · RTTSmAP: 45.47 -> 49.48Improved
VaT versus supervised AD-YOLO on synthetic hazy dataVaT · VOCh-testmAP: 84.11 -> 83.59No clear effect
VaT image quality versus unpaired methods on synthetic hazy dataVaT · VOCh-testPSNR: 27.871 -> 26.682Hurt

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