Musical Instrument Recognition Using Higher Order Moments

This paper presents a study on performance of western musical instrument recognition using higher order moments, such as skewness and kurtosis. Higher-order moments contain supplementary statistical information of the signal over the conventional first or second order moments. For example skewness is used for measuring the symmetry of the distribution of data and Kurtosis is measure of flatness of data. Higher-order moments provide segment level features. These segment level features are integrating with low level features like MFCC, Spectral features, Temporal features.
Experimental result shows that low level features integrated using higher order moments improve recognition accuracy by more than 8%. Accuracy of Musical Instrument Recognition is analysed using different combination of hybrid feature set combined with higher order moment.