AI Paper Atlas

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

Self-Supervised Character-to-Character Distillation for Text Recognition

Tongkun Guan, Wei Shen, Xue Yang, Qi Feng, Zekun Jiang, Xiaokang Yang

ICCV 2023 · Label-efficient and robust learning › Self-supervised and contrastive · approach: Invariance-based self-supervised learning · 50 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 character-to-character distillation instead of sequence-level DiGViT-Base · STR avg, ARD fine-tuningaverage accuracy: 94.92 -> 96.30Improved
use CCD instead of DiG with STD fine-tuningViT-Small · STR avg, STD fine-tuningaverage accuracy: 93.23 -> 93.59Improved
add distillation strategy (ablation)CCD · STR avg, ARD fine-tuningaverage accuracy: 91.45 -> 95.57Improved
use geometric plus color augmentation on irregular view instead of color-only (R2R) (ablation)CCD · STR avg, ARD fine-tuningaverage accuracy: 95.08 -> 95.57Improved
train a self-supervised segmentation network instead of using K-means masks directly (ablation)CCD · STR avg, ARD fine-tuningaverage accuracy: 95.33 -> 95.57Improved
select pseudo-label cluster using border heuristic instead of arbitrary cluster (ablation)text segmentation pseudo-labels · TextSegIoU (%): 40.1 -> 70.0Improved
train segmentation network on K-means pseudo-labels (ablation)text segmentation · TextSegIoU (%): 70.0 -> 73.6Improved
use CCD pretraining instead of DiG for text segmentationViT-Small · TextSegIoU (%): 83.1 -> 84.8Improved
use CCD pretraining instead of DiG for super-resolutionViT-Small · TextZoomaverage SSIM: 0.7522 -> 0.7843Improved
evaluate frozen encoder features: CCD vs DiGViT-Small · 11 STR benchmarks, frozen encoderaverage accuracy: 71.1 -> 69.9Hurt
use CCD with 10% label fraction instead of DiGViT-Small · 10% of ARDaverage accuracy: 73.5 -> 78.2Improved
use CCD with 1% label fraction instead of DiGViT-Small · 1% of ARDaverage accuracy: 62.9 -> 63.7Improved

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