AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

A Token-level Text Image Foundation Model for Document Understanding

Tongkun Guan, Zining Wang, Pei Fu, Zhengtao Guo, Wei Shen, Kai Zhou, Tiezhu Yue, Chen Duan, Hao Sun, Qianyi Jiang, Junfeng Luo, Xiaokang Yang

ICCV 2025 · Vision-language and multimodal › Multimodal LLMs · approach: Interleaved image-text generation with MLLMs · 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 token alignment at LLM level (ablation)TokenVL · Full-image text recognition TotalTextedit distance: 0.3592 -> 0.3547Improved
add token alignment at LLM level (ablation)TokenVL · Full-image text recognition IIT-CDIPedit distance: 0.2388 -> 0.1921Improved
add token alignment at LLM level (ablation)TokenVL · Full-image text recognition DocGenomeedit distance: 0.2837 -> 0.2806Improved
add token alignment at LLM level (ablation)TokenVL · InfoVQAANLS: 75.3 -> 76.5Improved
add token alignment at LLM level (ablation)TokenVL · ChartQAaccuracy: 86.5 -> 86.6No clear effect
add token alignment at LLM level (ablation)TokenVL · DocVQAANLS: 93.8 -> 94.2Improved
add token abstractor instead of simple pooling (ablation)TokenVL · DocVQAANLS: 93.1 -> 93.8Improved
add token abstractor instead of simple pooling (ablation)TokenVL · ChartQAaccuracy: 86.5 -> 86.5No clear effect
replace CLIP-L with TokenFD, zero-shot segmentationtext segmentation · zero-shot average over TextSeg, TotalText, HierTextaverage: 15.55 -> 32.61Improved
increase TokenFD input resolution from 448px to 1024pxTokenFD zero-shot segmentation · zero-shot averageaverage: 32.61 -> 34.59Improved
increase CLIP-L resolution 336px to 1024pxCLIP-L zero-shot segmentation · zero-shot averageaverage: 15.55 -> 15.81No clear effect
replace InternViT2.5 with TokenFD as frozen VFMVicuna-7B VQA · four VQA benchmarks averageANLS average: 71.25 -> 72.21Improved

Earlier methods this paper compares with

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