AI Paper Atlas

Research map / Image generation / GANs

Discriminative Region Proposal Adversarial Networks for High-Quality Image-to-Image Translation

Chao Wang, Haiyong Zheng, Zhibin Yu, Ziqiang Zheng, Zhaorui Gu, Bing Zheng

ECCV 2018 · Image generation › GANs · approach: Large-scale GAN latent analysis · 62 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 fake-mask operation (ablation)DRPAN · Cityscapes real to semantic labelsClass IOU: 0.39 -> 0.43Improved
add fake-mask operation (ablation)DRPAN · Cityscapes real to semantic labelsper-class accuracy: 0.48 -> 0.52Improved
add fake-mask operation (ablation)DRPAN · Cityscapes semantic labels to photo (FCN-8s score)per-class accuracy: 0.22 -> 0.24Improved
add fake-mask operation (ablation)DRPAN · Cityscapes semantic labels to photo (AMT)% Turkers labeled real: 14.3% -> 18.2%Improved
add fake-mask operation (ablation)DRPAN · Maps to aerial (AMT)% Turkers labeled real: 31.7% -> 33.4%Improved
add fake-mask operation (ablation)DRPAN · Cityscapes semantic labels to photo (FCN-8s score)Class IOU: 0.19 -> 0.19No clear effect
add fake-mask operation (ablation)DRPAN (w/o mask) · Single image de-rainingPSNR: 25.51 -> 25.87Improved
shrink discriminative region from 128 to 16 (ablation)DRPAN · Single image de-rainingPSNR: 25.87 -> 26.20Improved
shrink discriminative region from 16 to 32 (ablation)DRPAN · Single image de-rainingSSIM: 0.8712 -> 0.8788Improved
add reviser and L1 loss to continued PatchGAN training (ablation)PatchGAN · Facades (qualitative)visual artifactsImproved
continue training PatchGAN with L1 loss on discriminative region (ablation)PatchGAN · Facades (qualitative)visual artifactsNo clear effect
DRPAN instead of Pix2pixPix2pix · Cityscapes real to semantic labelsper-pixel accuracy: 0.83 -> 0.88Improved

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