期刊名称:Journal of Theoretical and Applied Information Technology
印刷版ISSN:1992-8645
电子版ISSN:1817-3195
出版年度:2018
卷号:96
期号:24
出版社:Journal of Theoretical and Applied
摘要:Automatic spoken language recognition refers to a sequence of processes aim to transfer human perception ability of identifying spoken languages to machine using computer program. In spite of great achievements in the domain, the task is still challenging to be practically efficient and reliable. This paper run throughout decades of research attempts approaching optimal languages identification accuracy comparable to human ability of identifying spoken language. Analysis methodologies of extracting most relevant speech information were reviewed. Achievements of approach based on language dependent linguistics rules and those based only on spectral attributes conveys in speech signal were investigated and compared. Exists of standard multilingual speech corpora offers evaluation and comparison of varies speech analysis methods and classification algorithms in single speech variability effects environment. In spite of great achievements, this demanding multilingual communities' communication solution, still looking flexible model easily accepting new language, shorten recognition time, overcoming difficulties of dialects and accents variations and mixed languages speech recognition.
关键词:language identification; features extraction; phonotactics; acoustic; mixed speech