AI Paper Atlas

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

Visual Text Generation in the Wild

Yuanzhi Zhu, Jiawei Liu, Feiyu Gao, Wenyu Liu, Xinggang Wang, Peng Wang, Fei Huang, Cong Yao, Zhibo Yang

ECCV 2024 · Image generation › Text-to-image diffusion · approach: Personalized and region-controlled generation · 13 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
use MLLM-planned reasonable regions and contents instead of SynthText layoutscene text synthesis · SceneVTG-benchmarkIoU: 7.31 -> 33.95Improved
fine-tune visual encoder in TRCG (ablation)TRCG · SceneVTG-benchmarkPD-Edge (x1e-4): 4.80 -> 0.68Improved
split region and content generation into two steps (ablation)TRCG · SceneVTG-benchmarkIoU: 7.97 -> 33.95Improved
add region masks to LVTR (ablation)LVTR · SceneVTG-benchmarkFID-R: 40.75 -> 33.34Improved
add foreground loss L_fore (ablation)LVTR · SceneVTG-benchmarkFID-R: 39.66 -> 33.34Improved
add background loss L_back (ablation)LVTR · SceneVTG-benchmarkFID-R: 36.96 -> 33.34Improved
vary lambda_f from 0.01 to 0.003 (ablation)LVTR · SceneVTG-benchmarkFID-R: 33.34 -> 32.75No clear effect
increase lambda_f to 0.03 (ablation)LVTR · SceneVTG-benchmarkFID-R: 33.34 -> 38.01Hurt
train OCR with SceneVTG data instead of diffusion-based text generatorsCRNN recognizer · Irregular text recognitionline accuracy: 26.08 -> 35.50 (vs SynthText)Improved
SceneVTG vs TextDiffuser on end-to-end F-scoreSceneVTG · SceneVTG-benchmark end-to-endF-score: 53.43 -> 52.03Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 2 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