Application in Extraction of Voice Recognition Characteristic for Relief Algorithm

Abstract:

Pertaining to the study on voice recognition issue, the characteristic selection study on voice signal is described via Relief Algorithm in this paper, it is started from 24-dimensional MFCC parameters to find out the most important MFCC parameters in the voice signal, under the condition that the recognition ratio doesnt change greatly, the optimized combination is performed for 24-dimensional MFCC parameters to provide a new orientation for characteristic of voice recognition.

• Periodical: Advanced Materials Research (Volumes 765-767)
• Pages: 2772-2775
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.765-767.2772
• Online since: September 2013
• Authors: Lun Gao, Tai Fu Li, Feng Wen
• Keywords: Relief, Voice Recognition Characteristic Selection

References

[1] Yu Jianchao, Zhang Ruilin. Speaker Recognition based on MFCC and LPCC [J]. Computer Engineering and Design, Periodical 05, 2009 1189-1191.

[2] Li Ze, Cui Xuan, Ma Yuting, Chen Junyu. Study on MFCC and LPCC Characteristic Parameters in the Speaker Recognition [J]. Journal of Henan Institute of Engineering (Natural Science Version), Periodical 02, 2010, 51-55.

[3] Zhang Xichun, Cao Yan, Zhang Jun, Wei Gang. Extraction Algorithm based on Voice PLP Characteristic for WDFT [J]. Pattern Recognition and Artificial Intelligence Volume 18, 2005, Periodical 4, 469-473.

[4] Kang Yan, Zhang Xueying Study on Voice Emotion Recognition based on ZCPA Parameters [J]. Shanxi Electronic Technology Periodical 3, 2011, 80-82.

[5] Liang Wuzhou, Zhang Xueying. Noise- Resistance Voice Recognition based on Over Zero Peak Amplitude Characteristics for Weight Combination [J]. Journal of Taiyuan University of Technology, Periodical 3, 2011, 84-86.