.png){kind=icon}
返回
MAGIC: Multi-granularity domain adaptation for text recognition
.png){kind=icon}
PRE
AI
DOI：10.1016/j.patcog.2024.111229.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Domain gaps between synthetic text and real-world text restrict current text recognition methods. One solution is to align features through Unsupervised Domain Adaptation (UDA). Most existing UDA-based text recognition methods extract global and local features to alleviate domain differences, only focusing on character distribution gaps. However, notable distribution gaps in character combinations exert a pivotal influence diverse text recognition tasks. To this end, we propose a Multi-level And multi-Granularity domain adaptation with entropy loss guIded text reCognition model, named MAGIC. It integrates Global-level Domain Adaptation (GDA) to mitigate image-level domain drift and Local-level Multi-granularity Domain Adaptation (LMDA) local feature shifts. Particularly, we design a subword-level domain discriminator to align the subword features relating to each character combination. Moreover, multi-granularity entropy minimization is used to optimize the target domain data for better domain adaptation. Experimental results on several types of text datasets demonstrate the effectiveness of MAGIC.
Keyword:
Text recognition
Unsupervised domain adaptation
Entropy minimization
Multi-granularity prediction
期刊
IF:
7.6
论文数:
1.3W
被引数:
4.5W
机构
引用论文
The role and effectiveness of climate commissions in engaging the public on climate change in the UK
OAAI
OAAI分享收藏Temperature proxy data and their significance for the understanding of pyroclastic density currents.png)
GeologyOAAI
GeologyIF0
OAAI分享收藏
PageNet: Towards End-to-End Weakly Supervised Page-Level Handwritten Chinese Text RecognitionPageNet: 迈向端到端弱监督页面级手写中文文本识别
PREAI
分享收藏
查看更多内容