Paper Info
Title
Deep Scattering Spectrum
Abstract
A scattering transform defines a locally translation invariant representation which is stable to time-warping deformation. It extends MFCC representations by computing modulation spectrum coefficients of multiple orders, through cascades of wavelet convolutions and modulus operators. Second-order scattering coefficients characterize transient phenomena such as attacks and amplitude modulation. A frequency transposition invariant representation is obtained by applying a scattering transform along log-frequency. State-the-of-art classification results are obtained for musical genre and phone classification on GTZAN and TIMIT databases, respectively.
Year
DOI
Venue
2013
10.1109/TSP.2014.2326991
Signal Processing, IEEE Transactions  
Keywords
DocType
Volume
acoustic wave scattering,amplitude modulation,audio signal processing,cepstral analysis,signal classification,signal representation,GTZAN database,MFCC,TIMIT database,audio classification,deep scattering spectrum,frequency transposition invariant representation,mel-frequency cepstral coefficients,modulus operators,musical genre,phone classification,scattering transform,second-order scattering coefficients,spectrum coefficients,time-warping deformation,transient phenomena,wavelet convolutions,Audio classification,MFCC,deep neural networks,modulation spectrum,wavelets
Journal
62
Issue
ISSN
Citations 
16
1053-587X
14
PageRank 
References 
Authors
0.88
36
2
Name
Order
Citations
PageRank
Joakim Andén1647.70
Stéphane Mallat24107718.30