AI Paper Atlas

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

GlyphControl: Glyph Conditional Control for Visual Text Generation

Yukang Yang, Dongnan Gui, Yuhui Yuan, Weicong Liang, Haisong Ding, Han Hu, Kai Chen

NeurIPS 2023 · Image generation › Text-to-image diffusion · approach: Layout-to-image control · 17 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 glyph-image spatial control using a ControlNet branchStable Diffusion 2.0-base · SimpleBenchAcc-hat (%): 33 -> 48Improved
Train on the larger LAION-Glyph dataset, from 100K to 10MGlyphControl · SimpleBenchAcc-hat (%): 37 -> 48Improved
Use small glyph font size instead of medium (ablation)GlyphControl-100K · SimpleBenchAcc (%): 30 -> 5Hurt
Use large glyph font size instead of medium (ablation)GlyphControl-100K · SimpleBenchAcc (%): 30 -> 23Hurt
Unfreeze and fine-tune the U-Net decoder (ablation)GlyphControl-100K fine-tuned on TextCaps 5K · TextCaps 5K fine-tuningAcc (%): 48 -> 61Improved
Increase TextCaps 5K fine-tuning from pre-trained to 10 epochs (ablation)GlyphControl-100K · TextCaps 5K fine-tuningAcc (%): 30 -> 48Improved
Increase TextCaps 5K fine-tuning from 10 to 20 epochs (ablation)GlyphControl-100K · TextCaps 5K fine-tuningAcc (%): 48 -> 50Improved
Increase TextCaps 5K fine-tuning from 20 to 40 epochs (ablation)GlyphControl-100K · TextCaps 5K fine-tuningAcc (%): 50 -> 61Improved
Generate many groups of small text in paragraphs (ablation)GlyphControl · Paragraphs with numerous small OCR boxesOCR readabilityHurt

Later papers that compared with this method

1 later accepted paper reports GlyphControl-10M as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.

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