Joint time-frequency scattering for audio classification

Joakim Anden, Vincent Lostanlen, Stephane Mallat
2015 2015 IEEE 25th International Workshop on Machine Learning for Signal Processing (MLSP)  
We introduce the joint time-frequency scattering transform, a time shift invariant descriptor of time-frequency structure for audio classification. It is obtained by applying a two-dimensional wavelet transform in time and log-frequency to a time-frequency wavelet scalogram. We show that this descriptor successfully characterizes complex time-frequency phenomena such as time-varying filters and frequency modulated excitations. State-of-the-art results are achieved for signal reconstruction and phone segment classification on the TIMIT dataset.
doi:10.1109/mlsp.2015.7324385 dblp:conf/mlsp/AndenLM15 fatcat:d5z4waomnjcqjh47njrhl6fuhm