文献类型：期刊文献

英文题名：Research on machine learning text recognition method based on Artificial Intelligence

作者：Nan, Zhu[1]

第一作者：Nan, Zhu

通讯作者：Nan, Z[1]

机构：[1]Xinxiang Univ, Comp & Informat Engn Coll, Xinxiang 453000, Henan, Peoples R China

第一机构：新乡学院计算机与信息工程学院

通讯机构：[1]Xinxiang Univ, Comp & Informat Engn Coll, Xinxiang 453000, Henan, Peoples R China.|[1107118]新乡学院计算机与信息工程学院;[11071]新乡学院;

年份：2018

卷号：21

期号：2

外文期刊名：JOURNAL OF ADVANCED OXIDATION TECHNOLOGIES

收录：;Scopus(收录号：2-s2.0-85051324921);WOS:【SCI-EXPANDED(收录号:WOS:000433154904139)】；

语种：英文

外文关键词：Artificial intelligence; Machine learning; Character recognition; Neural network; Method research

摘要：Machine learning is a new hotspot in the field of artificial intelligence, has been in machine vision, speech recognition, Natural Language Processing search, recommendation system, the network has been successfully applied to the intelligent robot has been successfully applied in complex system. This paper discusses the relationship between the classification method of feature extraction and recognition and integration and classification are analyzed, and then according to the basic idea of comprehensive integrated method, according to the characteristics of the typical Chinese characters character sets, put forward the corresponding recognition and integration method. On this basis, a printed Chinese character recognition system is established. In this paper, an improved network ensemble method is proposed, and three minimum distance classifiers are proposed to extract local features, and a complete recognition system is formed by this method. The test results show that the comprehensive recognition rate is higher than the original classifier, which proves the effectiveness of the method.