Login | Register

Acoustic anomaly detection using robust statistical energy processing

Title:

Acoustic anomaly detection using robust statistical energy processing

Salik, Farakh Nayaab John (2007) Acoustic anomaly detection using robust statistical energy processing. Masters thesis, Concordia University.

[thumbnail of MR40896.pdf]
Preview
Text (application/pdf)
MR40896.pdf - Accepted Version
4MB

Abstract

An anomaly is the specific event that causes the violation of a process observer's expectations about the process under observation. In this work, the problem of spatially locating an acoustic anomaly is addressed. Once reduced to a problem in robust statistics, an automated observer is designed to detect when high energy sources are introduced into an acoustic scene. Accounting for potential energy from signal amplitude, and kinetic energy from signal frequency in wavelet-filtered sub-bands, an outlier a robust statistical characterization scheme was developed using the Teager energy operator. With a statistical expectation of energy content in sub-bands, a methodology is designed to detect signal energies that violate the statistical expectation. These minor anomalies provide some sense that a fundamental change in energy has occurred in the sub-band. By examining how the signal is changing across all sub-bands, a detector is designed that is able to determine when a fundamental change occurs in the sub-band signal trends. Minor anomalies occurring during such changes are labeled as major anomalies. Using established localization methods, position estimates are obtained for the major anomalies in each sub-band. Accounting for the possibility of a source with spatiotemporal properties, the median of sub-band position estimates provides the final spatial information about the source.

Divisions:Concordia University > Gina Cody School of Engineering and Computer Science > Electrical and Computer Engineering
Item Type:Thesis (Masters)
Authors:Salik, Farakh Nayaab John
Pagination:xviii, 198 leaves : ill. ; 29 cm.
Institution:Concordia University
Degree Name:M.A. Sc.
Program:Electrical and Computer Engineering
Date:2007
Thesis Supervisor(s):Kahrizi, Mojtaba
Identification Number:LE 3 C66E44M 2007 S255
ID Code:975595
Deposited By: Concordia University Library
Deposited On:22 Jan 2013 16:11
Last Modified:13 Jul 2020 20:08
Related URLs:
All items in Spectrum are protected by copyright, with all rights reserved. The use of items is governed by Spectrum's terms of access.

Repository Staff Only: item control page

Downloads per month over past year

Research related to the current document (at the CORE website)
- Research related to the current document (at the CORE website)
Back to top Back to top