论文信息 - STRUCTURED SPARSITY FOR AUDIO SIGNALS

STRUCTURED SPARSITY FOR AUDIO SIGNALS

Regression problems with mixed-norm priors on time-frequency coefficients lead to structured, sparse representations of audio signals. In this contribution, a systematic formulation of thresholding operators that allow for weighting in the time-frequency domain is presented. The related iterative algorithms are then evaluated on synthetic and real-life audio signals in the context of denoising and multi-layer decomposition. Further, initial results on the influence of the shape of the weighting masks are presented.

Kai Siedenburg | M. Dörfler

[1] D. Donoho,et al. Atomic Decomposition by Basis Pursuit , 2001 .

[2] Monika Dörfler,et al. Time-Frequency Analysis for Music Signals: A Mathematical Approach , 2001 .

[3] M. Yuan,et al. Model selection and estimation in regression with grouped variables , 2006 .

[4] Stéphane Mallat,et al. A Wavelet Tour of Signal Processing - The Sparse Way, 3rd Edition , 2008 .

[5] I. Loris. On the performance of algorithms for the minimization of ℓ1-penalized functionals , 2007, 0710.4082.

[6] Marc Teboulle,et al. A Fast Iterative Shrinkage-Thresholding Algorithm for Linear Inverse Problems , 2009, SIAM J. Imaging Sci..

[7] Bruno Torrésani,et al. Sparsity and persistence: mixed norms provide simple signal models with dependent coefficients , 2009, Signal Image Video Process..

[8] M. Kowalski. Sparse regression using mixed norms , 2009 .