Noise Robust Speaker Identification using Reliable Sub-Band Selection in Multi-Band Approach

신뢰성 높은 서브밴드 선택을 이용한 잡음에 강인한 화자식별

  • Kim, Sung-Tak (School of Engineering, Information and Communications University) ;
  • Ji, Mi-Gyeong (School of Engineering, Information and Communications University) ;
  • Kim, Hoi-Rin (School of Engineering, Information and Communications University)
  • 김성탁 (한국정보통신대학교 공학부) ;
  • 지미경 (한국정보통신대학교 공학부) ;
  • 김희린 (한국정보통신대학교 공학부)
  • Published : 2007.05.18

Abstract

The conventional feature recombination technique is very effective in the band-limited noise condition, but in broad-band noise condition, the conventional feature recombination technique does not produce notable performance improvement compared with the full-band system. To cope with this drawback, we introduce a new technique of sub-band likelihood computation in the feature recombination, and propose a new feature recombination method by using this sub-band likelihood computation. Furthermore, the reliable sub-band selection based on the signal-to-noise ratio is used to improve the performance of this proposed feature recombination. Experimental results shows that the average error reduction rate in various noise condition is more than 27% compared with the conventional full-band speaker identification system.

Keywords