{"title":"Automatic speech recognition for connected words using DTW/HMM for English/ Hindi languages","authors":"S. Singhal, R. Dubey","doi":"10.1109/CCINTELS.2015.7437908","DOIUrl":null,"url":null,"abstract":"This work presents an automatic speech recognition (ASR) system for connected words. A connected ASR system has been implemented by extending an isolated word recognizer for speaker dependent data. The work has been applied for English as well as Hindi language. The traditional approach of Mel frequency cepsral coefficient (MFCC) is used as features of the speech signal. Hidden markov model (HMM) and dynamic time warping (DTW) are used at back-end for feature mapping of unknown utterances. A database of isolated English/Hindi words is created for training phase while sentences are used for testing phase. The results are expressed in terms of percentage word error rate (WER). The performance of system for two feature extraction techniques (HMM, DTW) is compared.","PeriodicalId":131816,"journal":{"name":"2015 Communication, Control and Intelligent Systems (CCIS)","volume":"20 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2015-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"11","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2015 Communication, Control and Intelligent Systems (CCIS)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/CCINTELS.2015.7437908","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 11
Abstract
This work presents an automatic speech recognition (ASR) system for connected words. A connected ASR system has been implemented by extending an isolated word recognizer for speaker dependent data. The work has been applied for English as well as Hindi language. The traditional approach of Mel frequency cepsral coefficient (MFCC) is used as features of the speech signal. Hidden markov model (HMM) and dynamic time warping (DTW) are used at back-end for feature mapping of unknown utterances. A database of isolated English/Hindi words is created for training phase while sentences are used for testing phase. The results are expressed in terms of percentage word error rate (WER). The performance of system for two feature extraction techniques (HMM, DTW) is compared.
京公网安备 11010802042870号
京ICP备2023020795号-1
分享
求助内容：
应助结果提醒方式：
扫码关注我们
