基于CapsNet的汉字字形表征模型

全文链接:10.16451/J.CNKI.ISSN1003-6059.201902009

CNKI 提供的 DOI 服务无响应,全文链接变更到期刊网页


基于CapsNet的汉字字形表征模型

谢海闻, 叶东毅, 陈昭炯
(福州大学 数学与计算机科学学院 福州 350108)

CapsNet-Based Chinese Character Font Representation Model

XIE Haiwen, YE Dongyi, CHEN Zhaojiong
(College of Mathematics and Computer Science, Fuzhou University, Fuzhou 350108)

摘要 提出基于胶囊神经网络(CapsNet)的汉字字形表征模型,通过表征汉字字形中的部件实现汉字字形的表征.首先,对任一汉字字形生成所有部件类别的表征向量.然后,根据部件存在概率,利用基于欧氏距离的离群点检测,选取相应的部件表征向量.最后,由选出的部件表征向量组成该汉字的字形表征.实验表明,文中模型在仅经过部件字形训练的情况下,即可有效识别汉字部件,同时自动生成汉字字形的有效表征.
关键词: 汉字字形 ; 胶囊神经网络(CapsNet) ; 表征模型 ; 部件识别 ; 汉字字形重构

Abstract:A CapsNet-based Chinese character font representation model is proposed to represent Chinese character font by the representation of components. Firstly, representative vectors of all categories are generated by the model. Then, a group of component representative vectors are selected by the Euclidean-distance-based outlier detection according to component probabilities. Finally, these vectors are utilized to form the Chinese character font representations. The experimental results show that the proposed model, merely trained on component fonts, is capable of identifying components of Chinese characters and automatically generating effective representation of Chinese characters.
Key words: Chinese Character Font ; Network(CapsNet) ; Representation Model ; Component Identification ; Chinese Character Font Reconstruction

基金资助: 国家自然科学基金项目(No.61672158)、福建省自然科学基金项目(No.2018J1798)、福建省高校产学合作项目(No.2018H6010)资助


引用本文:
谢海闻, 叶东毅, 陈昭炯. 基于CapsNet的汉字字形表征模型[J]. 模式识别与人工智能, 2019, 32(2): 169-176. XIE Haiwen, YE Dongyi, CHEN Zhaojiong. CapsNet-Based Chinese Character Font Representation Model. , 2019, 32(2): 169-176.

发表回复

您的电子邮箱地址不会被公开。 必填项已用*标注

此站点使用Akismet来减少垃圾评论。了解我们如何处理您的评论数据