Open AccessFeature selection based on MBFOA for audio signal classification under consideration of Gaussian white noise
Author(s) -
Arumugam Muthumari,
Kaliappan Mala
Publication year - 2018
Publication title -
iet signal processing
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.384
H-Index - 42
ISSN - 1751-9683
DOI - 10.1049/iet-spr.2016.0607
Subject(s) - speech recognition , audio signal , computer science , signal (programming language) , mel frequency cepstrum , pattern recognition (psychology) , feature selection , cepstrum , artificial intelligence , audio signal processing , feature (linguistics) , white noise , probabilistic logic , noise (video) , selection (genetic algorithm) , artificial neural network , feature extraction , feature vector , speech coding , telecommunications , linguistics , philosophy , image (mathematics) , programming language
Audio classification is a difficult task because of the issue of extracting and choosing the optimum audio features. To reduce the computational complication from existing methods, this study proposes a feature‐selection method based on modified bacterial foraging optimisation algorithm (MBFOA) for classification of audio signals. Enhanced mel‐frequency cepstral coefficient and enhanced power normalised cepstral coefficients with peak and pitch are estimated the signal feature and optimised using MBFOA with the fitness function. Using the probabilistic neural network, the audio signal is classified into music and speech signal. Then, if the signal is music, the signal is classified as cello, clarinet, flute etc. If the signal is detected as a speech, then it is again classified as male or female voice. This approach shows that it is possible to boost the classification accuracy by using different features and optimisation technique.