AI Paper Atlas

Research map / Label-efficient and robust learning / Self-supervised and contrastive

Appearance Discrepancy-guided Sequence Hybrid Masking for Robust Scene Text Recognition

Shihao Zou, Wei Wei, Leyang Xu, Kaihe Xu, Wenfeng Xie

AAAI 2026 · Label-efficient and robust learning › Self-supervised and contrastive · approach: Masked autoencoder pretraining · 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
use sequence hybrid masking instead of random masking (ablation)DSHM pre-training · Challenge sets avg (HOST, WOST, ArT, COCO, Uber), 5M pre-training imagesavg word accuracy: 75.5 -> 79.0Improved
add discrepancy guidance to random masking (HM vs RM) (ablation)DSHM pre-training · Challenge sets avg (HOST, WOST, ArT, COCO, Uber), 5M pre-training imagesavg word accuracy: 75.5 -> 77.2Improved
add sequence awareness (block expansion) to hybrid masking (ablation)DSHM pre-training · Challenge sets avg (HOST, WOST, ArT, COCO, Uber), 5M pre-training imagesavg word accuracy: 77.2 -> 79.0Improved
add discrepancy-weighted reconstruction loss (ablation)DSHM pre-training · Challenge sets avg (HOST, WOST, ArT, COCO, Uber), 5M pre-training imagesavg word accuracy: 77.3 -> 79.0Improved
add CLS consistency loss (ablation)DSHM pre-training · Challenge sets avg (HOST, WOST, ArT, COCO, Uber), 5M pre-training imagesavg word accuracy: 78.3 -> 79.0Improved
add discrepancy-conditioned token in decoder (ablation)DSHM pre-training · Challenge sets avg (HOST, WOST, ArT, COCO, Uber), 5M pre-training imagesavg word accuracy: 76.6 -> 79.0Improved
hard/random mask ratio 0.5/0.3 instead of 0.6/0.2 (ablation)DSHM pre-training · Challenge sets avg (HOST, WOST, ArT, COCO, Uber), 5M pre-training imagesavg word accuracy: 77.3 -> 79.0Improved
hard/random mask ratio 0.5/0.3 instead of 0.4/0.4 (ablation)DSHM pre-training · Challenge sets avg (HOST, WOST, ArT, COCO, Uber), 5M pre-training imagesavg word accuracy: 76.6 -> 79.0Improved
use DSHM pre-training instead of LMIMLMIM · Union14M benchmark (avg)word accuracy (%): 86.3 -> 87.0Improved
use DSHM pre-training instead of LMIMLMIM · Chinese benchmark (avg, with pre-training)accuracy (%): 82.2 -> 83.0Improved
use DSHM pre-training instead of LMIMLMIM · HOST (occluded/challenging)word accuracy (%): 69.6 -> 72.8Improved

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