潮汕话孤立字的语音识别算法设计与实现

北京理工大学珠海学院2020届本科生毕业设计The Design and Implementation of the Speech RecognitionAlgorithm of Isolated Words in Chaoshan DialectAbstractWith the rapid development of science and technology,humans have not stoppedto simply input information to the computer through the keyboard and mouse,andeagerly hope to communicate with computer and interact with the it in a simpler andfaster way.Under this expectation,the technology of computers to understand humanspeech came into being.This is the speech recognition technology.Speechrecognition technology has been applied in many fields,which has greatly changed allaspects of human life.In recent years,the protection and research of dialects have been highly valuedby the state,and many achievements have been made.As an important dialect inGuangdong,Chaoshan dialect has many users.At present,the research andapplication of Chaoshan dialect speech recognition is still rare.This article takes the promotion of hometown dialect and culture as the startingpoint,researches on Chaoshan dialect,combines speech processing technology,basedon TF2.0 deep learning framework,uses python as a programming language,anddevelops a Chaoshan dialect isolated speech recognition software.Using anauthoritative English corpus (about 2,000 corpora)to conduct an experimental test onthe designed algorithm,the recognition rate can reach 86%.Since there is no sharedcorpus of Chaoshan dialect yet,the author of this article recorded and organized asmall-scale speech corpus of isolated words of Chaoshan dialect.In the case ofinsufficient corpus data,the algorithm recognition rate still reached 70%.The paper reviews the development process of speech recognition technology,briefly introduces speech recognition technology,simplifies the description of eachstage of speech recognition,analyzes the principles of deep learning,and analyzes thespeech recognition of Chaoshan dialect based on deep learning Made a more detailedintroduction.Keywords:speech recognition,Chaoshan dialect,deep learning;