Noise-robust algorithm of speech features extraction for automatic speech recognition system

Considered problems of extracting acoustic features from speech signal. Proposed algorithm based on formants information. Achieved improvement compared to MFCC-PLP baseline for speech in noisy environment.