AI Paper Atlas

Research map / Label-efficient and robust learning / Knowledge distillation

Text Is Text, No Matter What: Unifying Text Recognition Using Knowledge Distillation

Ayan Kumar Bhunia, Aneeshan Sain, Pinaki Nath Chowdhury, Yi-Zhe Song

ICCV 2021 · Label-efficient and robust learning › Knowledge distillation · approach: Few-shot and metric distillation · 34 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 logits distillation loss (ablation)unified student · IC15word recognition accuracy (%): 70.4 -> 75.3Improved
add logits distillation loss (ablation)unified student · IAMword recognition accuracy (%): 81.8 -> 84.9Improved
add attention distillation loss (ablation)unified student · IC15word recognition accuracy (%): 75.3 -> 75.7Improved
add hint distillation loss (ablation)unified student · IC15word recognition accuracy (%): 75.7 -> 76.4Improved
add affinity distillation loss (ablation)unified student · IC15word recognition accuracy (%): 76.4 -> 76.9Improved
add affinity distillation loss (ablation)unified student · IAMword recognition accuracy (%): 85.9 -> 86.4Improved
add conditional distillation (ablation)unified student · IC15word recognition accuracy (%): reduced by 2.5% without itImproved
add conditional distillation (ablation)unified student · IAMword recognition accuracy (%): reduced by 0.4% without itImproved
vary omega from 1.01 to 1.10 (ablation)unified student · IC15word recognition accuracy (%): 76.8 / 76.9 / 76.9 / 76.8 / 76.8No clear effect
put hint loss on context vector instead of CNN features (ablation)unified student · IC15word recognition accuracy (%): 3.8% -> 5.1% improvementImproved
put hint loss on context vector instead of decoder hidden state (ablation)unified student · IC15word recognition accuracy (%): 4.6% -> 5.1% improvementImproved
add image rectification module to KD pipeline (ablation)unified student · IC15word recognition accuracy (%): 76.9 -> 75.9Hurt

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