AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Aggregation Cross-Entropy for Sequence Recognition

Zecheng Xie, Yaoxiong Huang, Yuanzhi Zhu, Lianwen Jin, Yuliang Liu, Lele Xie

CVPR 2019 · Architectures and efficiency › Attention and transformers · approach: State space and recurrent sequence models · 137 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
replace regression-based ACE with cross-entropy-based ACE (ablation)ACE (1D) with CRNN · IIIT5Kaccuracy: 19.4 -> 82.3Improved
replace regression-based ACE with cross-entropy-based ACE (ablation)ACE (1D) with CRNN · SVTaccuracy: 6.6 -> 82.6Improved
replace regression-based ACE with cross-entropy-based ACE (ablation)ACE (1D) with CRNN · IC03accuracy: 12.0 -> 92.1Improved
replace regression-based ACE with cross-entropy-based ACE (ablation)ACE (1D) with CRNN · IC13accuracy: 9.3 -> 89.7Improved
use ACE instead of the strongest listed baselineACE (2D) · SVT-P (50-word lexicon)accuracy: 94.3 -> 94.9Improved
use ACE instead of the strongest listed baselineACE (2D) · CUTEaccuracy: 76.8 -> 82.6Improved
use ACE instead of the strongest listed baselineACE (2D) · SVT-P (no lexicon)accuracy: 75.8 -> 70.1Hurt
use ACE with 2D predictionscene text recognition · SVT-P (full lexicon)accuracy: 83.7 -> 87.8Improved
use ACE instead of the strongest listed baselineACE (2D) · IC15accuracy: 68.2 -> 68.9Improved
use ACE instead of the strongest listed baselineACE (1D) · ICDAR 2013 HCTR (without LM)CR: 87.43 -> 91.68Improved
use ACE instead of the strongest listed baselineACE (1D) · ICDAR 2013 HCTR (without LM)AR: 88.79 -> 91.25Improved
use ACE instead of the strongest listed baselineACE (1D) · ICDAR 2013 HCTR (with LM)CR: 96.32 -> 96.70Improved

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