AI Paper Atlas

Research map / Video / Video-language understanding

VimTS: A Unified Video and Image Text Spotter for Enhancing the Cross-Domain Generalization

Yuliang Liu, Mingxin Huang, Hao Yan, Linger Deng, Weijia Wu, Hao Lü, Chunhua Shen, Lianwen Jin, Xiang Bai

TPAMI 2025 · Video › Video-language understanding · approach: Video-language pretraining · 6 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 PQGM to joint word+line full-tuning (ablation)VimTS tuning · TotalText-word -> IC15end-to-end H-mean (%): 62.5 -> 67.6Improved
add PQGM to joint word+line full-tuning (ablation)VimTS tuning · CTW1500-line -> TotalText-lineend-to-end H-mean (%): 36.6 -> 54.0Improved
add PQGM to joint word+line full-tuning (ablation)VimTS tuning · TotalText-word -> CTW1500-wordend-to-end H-mean (%): 71.1 -> 81.5Improved
naively merge word- and line-level data without PQGM (ablation)single-task baseline · CTW1500-line -> TotalText-lineend-to-end H-mean (%): 50.5 -> 36.6Hurt
replace full-tuning with task-aware adapter (3.0% params) (ablation)VimTS + PQGM · TotalText-word -> IC15end-to-end H-mean (%): 67.6 -> 68.0Improved
replace standard adapter with task-aware adapter (ablation)VimTS + PQGM · TotalText-word -> IC15end-to-end H-mean (%): 67.2 -> 68.0Improved
aggregate detection information before recognition in adapter (ablation)task-aware adapter · TotalText-word -> IC15end-to-end H-mean (%): 67.5 -> 68.0Improved
add synthetic VTD-368k to image-level training (ablation)VimTS · ICDAR2015 videoMOTA (%): 61.0 -> 61.8Improved
add real ICDAR2015 video data on top of VTD-368k (ablation)VimTS · ICDAR2015 videoMOTA (%): 61.8 -> 62.3Improved
add VTD-368k under stricter IoU 0.7 (ablation)VimTS · DSText v2MOTA (%): -14.4 -> -1.3Improved
add VTD-368k to TransDETR (ablation)TransDETR · ICDAR2015 videoID_F1 (%): 72.8 -> 74.0Improved
add VTD-368k to TransVTSpotter (ablation)TransVTSpotter · ICDAR2015 videoID_F1 (%): 61.1 -> 62.5Improved

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