Lazy Learning Based Segregation of Top-3 South Indian Languages with LSF-A Feature

Himadri Mukherjee, Moumita Dutta, Sk Md Obaidullah, Kc Santosh, Santanu Phadikar, Kaushik Roy

Abstract

Identification of language from voice signals is known as automatic language identification. It is very important for speech recognition in multi lingual countries like India where people use more than a single language while talking. Language identification for South Indian languages is difficult for a person without prior knowledge. In this paper, the top 3 most spoken languages of South India namely Telugu, Tamil and Kannada has been distinguished with the help of line spectral frequency based features namely LSF-A (Line spectral frequency-Approximation). Experiments have been performed on multiple datasets having as many as 21700 clips and a highest accuracy of 99.70% has been obtained with a lazy learning-based classifier.
BibTeX
@inproceedings{mukherjee2018lazy,
  author = {Himadri Mukherjee and Moumita Dutta and Sk Md Obaidullah and KC Santosh and Santanu Phadikar and Kaushik Roy},
  title = {{Lazy Learning Based Segregation of Top-3 South Indian Languages with LSF-A Feature}},
  booktitle = {RTIP2R},
  year = {2018},
  doi = {10.1007/978-981-13-9181-1\_40},
}

Loading related content...

Figure