Language identification usingGaussian mixture models

[摘要] ENGLISH ABSTRACT: The importance of Language Identification for African languages is seeing adramatic increase due to the development of telecommunication infrastructureand, as a result, an increase in volumes of data and speech traffic in publicnetworks. By automatically processing the raw speech data the vital assistancegiven to people in distress can be speeded up, by referring their calls to a personknowledgeable in that language.To this effect a speech corpus was developed and various algorithms were implementedand tested on raw telephone speech data. These algorithms entaileddata preparation, signal processing, and statistical analysis aimed at discriminatingbetween languages. The statistical model of Gaussian Mixture Models(GMMs) were chosen for this research due to their ability to represent an entirelanguage with a single stochastic model that does not require phonetic transcription.Language Identification for African languages using GMMs is feasible, althoughthere are some few challenges like proper classification and accuratestudy into the relationship of langauges that need to be overcome. Other methodsthat make use of phonetically transcribed data need to be explored andtested with the new corpus for the research to be more rigorous.

[发布机构] Stellenbosch University