AI Paper Atlas

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

ELDiff: When Evidential Learning Meets Text-to-Image Diffusion

Qingtao Pan, Kai Ye, Zhihao Dou, Bing Ji, Shuo Li

ECCV 2026 · Image generation › Text-to-image diffusion · approach: Compositional text-to-image guidance · 0 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 pixel evidence loss on top of plain LDM finetuning (ablation)SD v1.4 finetune · VISOR OAOA: 0.3821 -> 0.5466Improved
add token conflict loss to pixel evidence loss (ablation)SD v1.4 finetune · VISOR OAOA: 0.5466 -> 0.5563Improved
add token conflict loss to pixel evidence loss (ablation)SD v1.4 finetune · MultiGen MG2MG2 success rate: 0.9821 -> 0.9736Hurt
add token conflict loss to pixel evidence loss (ablation)SD v1.4 finetune · MultiGen MG5MG5 success rate: 0.3361 -> 0.3530Improved
add pixel evidence loss on top of plain LDM finetuning (ablation)SD v1.4 finetune · COCO instance valFID: 24.57 -> 21.17Improved
finetune with plain LDM loss only (vs frozen SD v1.4) (ablation)SD v1.4 · COCO instance valFID: 22.06 -> 24.57Hurt
ELDiff finetuning instead of TokenComposeSD v2.1 · VISOR OAOA: 0.6010 -> 0.7116Improved
ELDiff finetuning instead of TokenComposeSD v3.5 Medium · VISOR OAOA: 0.8233 -> 0.8644Improved
plain LDM finetuning (no consistency loss) vs frozenSD v3.5 Medium · VISOR OAOA: 0.8084 -> 0.8010Hurt
ELDiff finetuning vs frozen SDXLSDXL · T2I-CompBench Colorcolor attribute binding: 0.6734 -> 0.7768Improved
ELDiff finetuning vs base Qwen-ImageQwen-Image · GenEval Overalloverall: 0.87 -> 0.90Improved
ELDiff finetuning vs base Qwen-ImageQwen-Image · GenEval 2 Soft-TIFA GMSoft-TIFA GM: 34.57 -> 37.94Improved

Earlier methods this paper compares with

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