Time-frequency audio feature extraction based on tensor representation of sparse coding

Abstract

A time-frequency audio feature extraction scheme is proposed, in which features are decomposed from a frequency-time-scale tensor. The tensor, derived from a weight vector and a Gabor dictionary in sparse coding, represents the frequency, time centre and scale of transient time-frequency components with different dimensions. The distinguishing Gabor atoms are represented by individual tensor elements, and their associated coding weights are represented by tensor element values. The experimental results of sound effects classification showed performance improvement against that of sparse coding features.