Forensic Audio Analysis and Event Recognition for Smart Surveillance Systems
Authors:
- 1 November 2019
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Abstract
In this study, we propose forensic audio analysis and event recognition framework for smart surveillance systems. Surveillance systems are getting a growing attention. Integration of audio surveillance systems to IP cameras offers advantages over video surveillance systems solely. For instance, an audio surveillance system can provide an additional information about the exact moment of an incident that might be occurring out of the range of surveillance cameras. In our work, mel-frequency cepstral coefficients (MFCCs), energy, pitch range (PR), and linear prediction coefficients (LPC) feature sets were extracted and evaluated on an open-access DASE database, which contains nine audio events including glass breaking, dog barking, scream, gunshot, explosions, police sirens, door slams, footsteps, and house alarm sounds. K-nearest-neighbor (KNN), support vector machines (SVMs), Gaussian Mixture Model (GMM), and classifier fusion classifiers are designed. All classifiers achieved promising accuracy rates.Keywords
This publication has 17 references indexed in Scilit:
- Audio SurveillanceACM Computing Surveys, 2016
- A new pitch-range based feature set for a speaker’s age and gender classificationApplied Acoustics, 2015
- Application of abnormal sound recognition system for indoor environment2013
- On investigating efficient methodology for Environmental Sound Recognition2013
- Selective Background Adaptation Based Abnormal Acoustic Event Recognition for Audio Surveillance2012
- Environmental noise classifier using a new set of feature parameters based on pitch rangeApplied Acoustics, 2011
- LIBSVMACM Transactions on Intelligent Systems and Technology, 2011
- Environmental Sound Recognition With Time–Frequency Audio FeaturesIEEE Transactions on Audio, Speech, and Language Processing, 2009
- Improved One-class SVM Classifier for Sounds Classification2007
- Towards parameter-free classification of sound effects in movies2005