AI Paper Atlas

Research map / Image generation / GANs

Semantics-Enhanced Adversarial Nets for Text-to-Image Synthesis

Hongchen Tan, Xiuping Liu, Xin Li, Yi Zhang, Baocai Yin

ICCV 2019 · Image generation › GANs · approach: Large-scale GAN latent analysis · 93 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 attention competition module (AC) (ablation)AttnGAN · CUB testInception score: 4.31 -> 4.61Improved
add attention competition module (AC) (ablation)AttnGAN · COCO testInception score: 25.56 -> 26.83Improved
add Siamese network with sliding loss (SL) (ablation)AttnGAN · CUB testInception score: 4.31 -> 4.58Improved
add Siamese network with sliding loss (SL) (ablation)AttnGAN · COCO testInception score: 25.56 -> 27.13Improved
add Siamese network with contrastive loss (ablation)AttnGAN · CUB testInception score: 4.31 -> 4.44Improved
replace contrastive loss with sliding loss (gamma=1/2) (ablation)AttnGAN + SiaNet · CUB testInception score: 4.44 -> 4.58Improved
add both ACM and SCM (ablation)AttnGAN · CUB testInception score: 4.31 -> 4.67Improved
add both ACM and SCM (ablation)AttnGAN · COCO testInception score: 25.56 -> 27.86Improved
increase sliding loss gamma from 1/2 to 2 (ablation)AttnGAN + SL · CUB testInception score: 4.58 -> 4.41Hurt
set attention threshold alpha=1 instead of 0.005 (ablation)AttnGAN + AC · CUB testInception score: 4.61 -> 4.25Hurt
use alpha 0.005 instead of 0.001 (ablation)AttnGAN + AC · CUB testInception score: 4.41 -> 4.61Improved
add SEGAN modulesAttnGAN · CUB testFID: 22.504 -> 18.167Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 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