AI Paper Atlas

Research map / Image generation / Text-to-image diffusion

Getting it Right: Improving Spatial Consistency in Text-to-Image Models

Agneet Chatterjee, Gabriela Ben Melech Stan, Estelle Guez Aflalo, Sayak Paul, Dhruba Ghosh, Tejas Gokhale, Ludwig Schmidt, Hanna Hajishirzi, Vasudev Lal, Chitta R Baral, Yezhou Yang

ECCV 2024 · Image generation › Text-to-image diffusion · approach: T2I evaluation and controllable adapters · 21 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
fine-tune SD 2.1 on spatially focused SPRIGHT captions (<15k images)Stable Diffusion 2.1 · T2I-CompBench spatialspatial score: 0.1507 -> 0.1840Improved
fine-tune SD 2.1 on SPRIGHT captions (<15k images)Stable Diffusion 2.1 · COCO-30KZS-FID: 21.646 -> 14.925Improved
train only on images with more than 18 objects (444 images) (ablation)Stable Diffusion 2.1 + SPRIGHT · T2I-CompBench spatialspatial score: 0.1309 -> 0.2133Improved
use 50% spatial captions instead of 100% spatial captions in fine-tuning mix (ablation)Stable Diffusion 2.1 · T2I-CompBench spatialspatial score: 0.140 -> 0.178Improved
use only 25% spatial captions instead of 50% (ablation)Stable Diffusion 2.1 · T2I-CompBench spatialspatial score: 0.178 -> 0.154Hurt
use long spatial captions instead of short single-sentence captions (ablation)Stable Diffusion 2.1 with CLIP fine-tuning · T2I-CompBench spatialspatial score: 0.1230 -> 0.1777Improved
fine-tune CLIP text encoder together with U-Net (ablation)Stable Diffusion 2.1 with long captions · T2I-CompBench spatialspatial score: 0.1605 -> 0.1777Improved
fine-tune CLIP text encoder with short spatial captions (ablation)Stable Diffusion 2.1 · T2I-CompBench spatialspatial score: 0.1420 -> 0.1230Hurt
replace spatial phrases with negations during training (ablation)Stable Diffusion 2.1 + SPRIGHT · T2I-CompBench spatial (negation prompts)spatial score: 0.066 -> 0.069 (negation only); 0.1376 -> 0.1427 (mixed)No clear effect
fine-tune on SPRIGHT, effect on GenEval attribute bindingStable Diffusion 2.1 · GenEvalattribute binding: 0.17 -> 0.15Hurt
fine-tune on SPRIGHT, effect on GenEval counting and positionStable Diffusion 2.1 · GenEvalcounting / position: 0.44 -> 0.49 (counting); 0.07 -> 0.11 (position)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