A Causal Locally Competitive Algorithm for the Sparse Decomposition of Audio Signals

Adam S. Charles; Abigail Anne Kressner; Christopher J. Rozell

doi:10.1109/DSP-SPE.2011.5739223

A Causal Locally Competitive Algorithm for the Sparse Decomposition of Audio Signals

Adam S. Charles, Abigail Anne Kressner, Christopher J. Rozell

Georgia Institute of Technology

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

Abstract

While current inference methods can decompose audio signals, they require the entire signal upfront and are therefore ill-suited for real-time applications requiring causal processing. We propose a neurally-inspired, causal, sparse inference scheme based on the Locally Competitive Algorithm (LCA) over a temporal-spectral neighborhood. We demonstrate that this causal inference scheme can achieve lower sparsity levels and better signal fidelity than current filter and threshold approaches. Additionally, for some regimes, the sparsity level approaches those of Matching Pursuit while still maintaining signal integrity.

Original language	English
Title of host publication	Proceedings of 2011 IEEE Digital Signal Processing Workshop and IEEE Signal Processing Education Workshop
Publisher	IEEE
Publication date	2011
Pages	265-270
ISBN (Print)	978-1-61284-226-4
DOIs	https://doi.org/10.1109/DSP-SPE.2011.5739223
Publication status	Published - 2011
Externally published	Yes
Event	2011 Digital Signal Processing and Signal Processing Education Meeting - Sedona, United States Duration: 4 Jan 2011 → 7 Jan 2011 https://ieeexplore.ieee.org/xpl/conhome/5731631/proceeding

Conference

Conference	2011 Digital Signal Processing and Signal Processing Education Meeting
Country/Territory	United States
City	Sedona
Period	04/01/2011 → 07/01/2011
Internet address	https://ieeexplore.ieee.org/xpl/conhome/5731631/proceeding

Keywords

ENGINEERING,
GAMMACHIRP
Locally Competitive Algorithm (LCA)
causal sparse encoding
audio processing
convolutional model

Access to Document

10.1109/DSP-SPE.2011.5739223

OpenUrl availability

Full text

Cite this

@inproceedings{e32a7aee8f984548b0aa933807b03514,

title = "A Causal Locally Competitive Algorithm for the Sparse Decomposition of Audio Signals",

abstract = "While current inference methods can decompose audio signals, they require the entire signal upfront and are therefore ill-suited for real-time applications requiring causal processing. We propose a neurally-inspired, causal, sparse inference scheme based on the Locally Competitive Algorithm (LCA) over a temporal-spectral neighborhood. We demonstrate that this causal inference scheme can achieve lower sparsity levels and better signal fidelity than current filter and threshold approaches. Additionally, for some regimes, the sparsity level approaches those of Matching Pursuit while still maintaining signal integrity.",

keywords = "ENGINEERING,, GAMMACHIRP, Locally Competitive Algorithm (LCA), causal sparse encoding, audio processing, convolutional model",

author = "Charles, {Adam S.} and Kressner, {Abigail Anne} and Rozell, {Christopher J.}",

year = "2011",

doi = "10.1109/DSP-SPE.2011.5739223",

language = "English",

isbn = "978-1-61284-226-4",

pages = "265--270",

booktitle = "Proceedings of 2011 IEEE Digital Signal Processing Workshop and IEEE Signal Processing Education Workshop",

publisher = "IEEE",

address = "United States",

note = "2011 Digital Signal Processing and Signal Processing Education Meeting, DSP/SPE 2011 ; Conference date: 04-01-2011 Through 07-01-2011",

url = "https://ieeexplore.ieee.org/xpl/conhome/5731631/proceeding",

}

Charles, AS, Kressner, AA & Rozell, CJ 2011, A Causal Locally Competitive Algorithm for the Sparse Decomposition of Audio Signals. in Proceedings of 2011 IEEE Digital Signal Processing Workshop and IEEE Signal Processing Education Workshop. IEEE, pp. 265-270, 2011 Digital Signal Processing and Signal Processing Education Meeting, Sedona, Arizona, United States, 04/01/2011. https://doi.org/10.1109/DSP-SPE.2011.5739223

A Causal Locally Competitive Algorithm for the Sparse Decomposition of Audio Signals. / Charles, Adam S.; Kressner, Abigail Anne; Rozell, Christopher J.
Proceedings of 2011 IEEE Digital Signal Processing Workshop and IEEE Signal Processing Education Workshop. IEEE, 2011. p. 265-270.

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

TY - GEN

T1 - A Causal Locally Competitive Algorithm for the Sparse Decomposition of Audio Signals

AU - Charles, Adam S.

AU - Kressner, Abigail Anne

AU - Rozell, Christopher J.

PY - 2011

Y1 - 2011

N2 - While current inference methods can decompose audio signals, they require the entire signal upfront and are therefore ill-suited for real-time applications requiring causal processing. We propose a neurally-inspired, causal, sparse inference scheme based on the Locally Competitive Algorithm (LCA) over a temporal-spectral neighborhood. We demonstrate that this causal inference scheme can achieve lower sparsity levels and better signal fidelity than current filter and threshold approaches. Additionally, for some regimes, the sparsity level approaches those of Matching Pursuit while still maintaining signal integrity.

AB - While current inference methods can decompose audio signals, they require the entire signal upfront and are therefore ill-suited for real-time applications requiring causal processing. We propose a neurally-inspired, causal, sparse inference scheme based on the Locally Competitive Algorithm (LCA) over a temporal-spectral neighborhood. We demonstrate that this causal inference scheme can achieve lower sparsity levels and better signal fidelity than current filter and threshold approaches. Additionally, for some regimes, the sparsity level approaches those of Matching Pursuit while still maintaining signal integrity.

KW - ENGINEERING,

KW - GAMMACHIRP

KW - Locally Competitive Algorithm (LCA)

KW - causal sparse encoding

KW - audio processing

KW - convolutional model

U2 - 10.1109/DSP-SPE.2011.5739223

DO - 10.1109/DSP-SPE.2011.5739223

M3 - Article in proceedings

SN - 978-1-61284-226-4

SP - 265

EP - 270

BT - Proceedings of 2011 IEEE Digital Signal Processing Workshop and IEEE Signal Processing Education Workshop

PB - IEEE

T2 - 2011 Digital Signal Processing and Signal Processing Education Meeting

Y2 - 4 January 2011 through 7 January 2011

ER -

A Causal Locally Competitive Algorithm for the Sparse Decomposition of Audio Signals

Abstract

Conference

Keywords

Access to Document

OpenUrl availability

Fingerprint

Cite this