基于CNN-SVM的离线手写新太略字符识别

2019 IEEE 2nd International Conference on Electronic Information and Communication Technology (ICEICT) Pub Date : 2019-01-01 DOI:10.1109/ICEICT.2019.8846292

Yongqiang Wang, Pengfei Yu

{"title":"基于CNN-SVM的离线手写新太略字符识别","authors":"Yongqiang Wang, Pengfei Yu","doi":"10.1109/ICEICT.2019.8846292","DOIUrl":null,"url":null,"abstract":"In this paper, the classical convolutional neural network VGG19 is used to recognize the offline handwritten New Tai Lue characters firstly. Then, according to the recognized image features, the CNN suitable for this data set is constructed repeatedly, and 83 kinds of offline handwritten characters are experimentally tested. Finally, the features extracted from CNN are classified by Support Vector Machine (SVM) and the recognition rate is improved compared with VGG19 and CNN constructed in this paper.","PeriodicalId":382686,"journal":{"name":"2019 IEEE 2nd International Conference on Electronic Information and Communication Technology (ICEICT)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2019-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"5","resultStr":"{\"title\":\"Offline Handwritten New Tai Lue Characters Recognition Using CNN-SVM\",\"authors\":\"Yongqiang Wang, Pengfei Yu\",\"doi\":\"10.1109/ICEICT.2019.8846292\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"In this paper, the classical convolutional neural network VGG19 is used to recognize the offline handwritten New Tai Lue characters firstly. Then, according to the recognized image features, the CNN suitable for this data set is constructed repeatedly, and 83 kinds of offline handwritten characters are experimentally tested. Finally, the features extracted from CNN are classified by Support Vector Machine (SVM) and the recognition rate is improved compared with VGG19 and CNN constructed in this paper.\",\"PeriodicalId\":382686,\"journal\":{\"name\":\"2019 IEEE 2nd International Conference on Electronic Information and Communication Technology (ICEICT)\",\"volume\":\"22 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2019-01-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"5\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2019 IEEE 2nd International Conference on Electronic Information and Communication Technology (ICEICT)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/ICEICT.2019.8846292\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2019 IEEE 2nd International Conference on Electronic Information and Communication Technology (ICEICT)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ICEICT.2019.8846292","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}

引用次数: 5

摘要

本文首先利用经典卷积神经网络VGG19对离线手写新太略汉字进行识别。然后，根据识别出的图像特征，重复构建适合该数据集的CNN，并对83种离线手写字符进行实验测试。最后，利用支持向量机(SVM)对CNN提取的特征进行分类，与本文构建的VGG19和CNN相比，提高了识别率。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

查看原文本刊更多论文

Offline Handwritten New Tai Lue Characters Recognition Using CNN-SVM

In this paper, the classical convolutional neural network VGG19 is used to recognize the offline handwritten New Tai Lue characters firstly. Then, according to the recognized image features, the CNN suitable for this data set is constructed repeatedly, and 83 kinds of offline handwritten characters are experimentally tested. Finally, the features extracted from CNN are classified by Support Vector Machine (SVM) and the recognition rate is improved compared with VGG19 and CNN constructed in this paper.

求助全文

通过发布文献求助，成功后即可免费获取论文全文。去求助

来源期刊

2019 IEEE 2nd International Conference on Electronic Information and Communication Technology (ICEICT)

自引率

0.00%

发文量