AI Paper Atlas

Research map / Label-efficient and robust learning / Domain adaptation and generalization

Sequence-To-Sequence Domain Adaptation Network for Robust Text Image Recognition

Yaping Zhang, Shuai Nie, Wenju Liu, Xing Xu, Dongxiang Zhang, Heng Tao Shen

CVPR 2019 · Label-efficient and robust learning › Domain adaptation and generalization · approach: Conditional distribution alignment · 152 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
Add gated attention similarity (GAS) to the sequence recognizer (ablation)SSDAN · IAM testWER: 23.9 -> 22.2Improved
Add gated attention similarity (GAS) to the sequence recognizer (ablation)SSDAN · CROHME 2014ExpRate: 39.9 -> 41.6Improved
Add gated attention similarity (GAS) to the sequence recognizer (ablation)SSDAN · IIIT5K, SVT, IC-03, and IC-13accuracy: 81.1 -> 83.8; 82.1 -> 84.5; 91.2 -> 92.1; 91.0 -> 91.8Improved
Align CNN outputs directly rather than attended character-level features (ablation)SSDAN on IAM · IAM testWER: 23.9 -> 25.0Hurt
Align CNN outputs directly rather than attended character-level features (ablation)SSDAN on IAM · IAM testCER: 9.2 -> 11.1Hurt
Use DenseNet instead of VGG or ResNet as the encoder (ablation)SSDAN on IAM · IAM testWER: 26.9 -> 22.2; 27.9 -> 22.2Improved
Use DenseNet instead of VGG or ResNet as the encoder (ablation)SSDAN on IAM · IAM testCER: 12.6 -> 8.5; 13.1 -> 8.5Improved
Use CORAL instead of MMD or adversarial loss for domain-shift measurement (ablation)SSDAN on IAM · IAM testWER: 22.7 -> 22.2; 24.6 -> 22.2Improved
Use CORAL instead of MMD or adversarial loss for domain-shift measurement (ablation)SSDAN on IAM · IAM testCER: 8.8 -> 8.5; 10.8 -> 8.5Improved
Vary the confidence threshold in the attention gate (ablation)SSDAN on MJSYNTH → SVT · SVTaccuracyMixed
Vary the domain-adaptation loss weight λ (ablation)SSDAN on MJSYNTH → SVT · SVTaccuracyMixed
Add unlabeled target examples alongside labeled training examples (ablation)SSDAN · IC-03 and IIIT5KaccuracyImproved

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