Title :
Feature extraction based on auditory representations for robust speech recognition
Author :
Kim, Doh-Suk ; Lee, Soo-Young ; Kil, Rhee M. ; Zhu, Xuelong
Author_Institution :
Dept. of Electr. Eng., Korea Adv. Inst. of Sci. & Technol., Taejon, South Korea
fDate :
1/2/1997 12:00:00 AM
Abstract :
A computational auditory model based on the detection of zero-crossings is introduced for robust speech recognition. The proposed method not only preserves intensity information, but is also robust to noise in frequency estimation. The robustness of the proposed method is verified analytically and experimentally
Keywords :
feature extraction; speech recognition; computational auditory model; feature extraction; frequency estimation; intensity information; noise; robust speech recognition; zero-crossings;
Journal_Title :
Electronics Letters
DOI :
10.1049/el:19970005
