AI Paper Atlas

Research map / Vision-language and multimodal / Human and embodied modeling

On Vocabulary Reliance in Scene Text Recognition

Zhaoyi Wan, Jielei Zhang, Liang Zhang, Jiebo Luo, Cong Yao

CVPR 2020 · Vision-language and multimodal › Human and embodied modeling · approach: Language-grounded scene text and structure · 75 citations (OpenAlex)

Check 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 mutual learning with a segmentation-based decoderAttention decoder · Scene text recognition evaluation frameworkvocabulary generalization (VG): 77.1 to 93.6Improved
Add mutual learning with an attention-based decoderSegmentation-based decoder · Scene text recognition evaluation frameworkvocabulary learning ability (VA): 75.8 to 82.3Improved
Mix RS data into LS training data at ratio 25% (ablation)Attention decoder without context module · IIITaccuracy: 77.8% to 84.4%Improved
Add PPM context modeling (ablation)Segmentation-based decoder · Scene text recognition evaluation frameworkvocabulary learning ability (VA): +3.9%Improved
Add PPM context modeling (ablation)CTC decoder · Scene text recognition evaluation frameworkvocabulary learning ability (VA): +10.5%Improved
Add context modeling (ablation)CTC-family models · Scene text recognition evaluation frameworkvocabulary generalization (VG)Hurt
Add BLSTM context modeling (ablation)Attention decoder · Scene text recognition evaluation frameworkgeneral accuracy (GA): 81.0 to 83.1Improved
Add mutual learning with an attention-based decoder (ablation)Segmentation-based decoder · Scene text recognition evaluation frameworkvocabulary generalization (VG): 97.3 to 96.0Hurt

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