ETRW on Speech Processing in Adverse Conditions Cannes-Mandelieu, France November 10-13, 1992

Robust Features for Applications in Speech and Speaker Recognition

B. Yegnanarayana, A. S. Madhukumar, V. R. Ramachandran

Department of Computer Science and Engineering, Indian Institute of Technology, Madras, India

This paper proposes the use of robust suprasegmental features for word boundary hypothesization in a speech to text system for Hindi and for a speaker identification system under clean as well as adverse speech input conditions. The features are derived from pitch accent patterns in Hindi.

Full Paper

Bibliographic reference.  Yegnanarayana, B. / Madhukumar, A. S. / Ramachandran, V. R. (1992): "Robust features for applications in speech and speaker recognition", In SPAC-1992, 97-101.