Noise robust speech recognition system using Mel cepstral and genetic algorithm
Mamta Garg, Arora Ajat Shatru, Gupta Savita · 2016
This paper suggested a technique based on MFCC analysis for audio signals with speech classification application. The proposed work used multi-resolution (wavelet) analysis and spectral analysis based features for feature extraction. The proposed approach uses a no. of features like Mel Frequency Cepstral Coefficient (MFCC), and FFT Coefficients combined with wavelet based features. In addition, accuracy of 99% using the proposed work is claimed.