AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

SCATTER: Selective Context Attentional Scene Text Recognizer

Ron Litman, Oron Anschel, Shahar Tsiper, Roee Litman, Shai Mazor, R. Manmatha

CVPR 2020 · Architectures and efficiency › Attention and transformers · approach: Long-context KV and memory attention · 169 citations (OpenAlex)

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 SCATTER with 5 blocks rather than Mask TextSpotterICDAR2015 recognition · IC15accuracy (%): 78.2% vs. 82.2%; +4.0 ppImproved
Use SCATTER with 5 blocks rather than Mask TextSpotterSVTP recognition · SVTPaccuracy (%): 83.6% vs. 86.9%; +3.3 ppImproved
Use SCATTER with 5 blocks rather than Mask TextSpotterCUTE recognition · CUTEaccuracy (%): 88.5% vs. 87.5%Hurt
Add intermediate CTC supervision (ablation)SCATTER 1-block model · Regular text weighted averageaccuracy (%): +0.2 ppImproved
Add intermediate CTC supervision (ablation)SCATTER 1-block model · Irregular text weighted averageaccuracy (%): +0.4 ppImproved
Replace the standard attentional decoder with the selective decoder (ablation)SCATTER 1-block model · Regular text weighted averageaccuracy (%): +0.4 ppImproved
Replace the standard attentional decoder with the selective decoder (ablation)SCATTER 1-block model · Irregular text weighted averageaccuracy (%): +2.7 ppImproved
Increase the number of intermediate selective decoders from 1 to 3 (ablation)SCATTER with 4 BiLSTM layers · Regular text weighted averageaccuracy (%): +0.7 ppImproved
Increase the number of intermediate selective decoders from 1 to 3 (ablation)SCATTER with 4 BiLSTM layers · Irregular text weighted averageaccuracy (%): +2.9 ppImproved
Add intermediate supervision and stack more BiLSTM layers (ablation)SCATTER encoder · Regular and irregular text weighted averagesaccuracy (%): 92.7 -> 94.0; 79.1 -> 83.7Improved
Increase encoder depth beyond 10 BiLSTM layers (ablation)SCATTER · Regular and irregular text weighted averagesaccuracy (%): -0.4 pp; -0.5 ppHurt
Train with 5 blocks, then prune to one block for inference (ablation)SCATTER · Regular text weighted averageaccuracy (%): +0.4 ppImproved

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