Zero-shot Handwritten Chinese Character Recognition with hierarchical decomposition embedding

作者：

Highlights：

• A novel hierarchical decomposing embedding (HDE) method is proposed.

• The framework HDE-Net is proposed for zero-shot learning.

• HDE-Net achieves state-of-the-art results on CASIA-HWDB, ICDAR, CTW datasets.

• Qualitative and quantitative analyses demonstrate the effectiveness of the proposed framework.

摘要

•A novel hierarchical decomposing embedding (HDE) method is proposed.•The framework HDE-Net is proposed for zero-shot learning.•HDE-Net achieves state-of-the-art results on CASIA-HWDB, ICDAR, CTW datasets.•Qualitative and quantitative analyses demonstrate the effectiveness of the proposed framework.

论文关键词：Chinese character recognition,Radical analysis,Zero-shot learning,Label embedding

论文评审过程：Received 5 November 2019, Revised 21 May 2020, Accepted 3 June 2020, Available online 6 June 2020, Version of Record 22 June 2020.

论文官网地址：https://doi.org/10.1016/j.patcog.2020.107488