Classification of audio sources using neural network applicable in security or military industry

• Main Authors: Navratil, M, Dostalek, P, Kresalek, V
• Format: Conference Proceeding
• Language: English
• Subjects: analysis; Artificial neural networks; audio; Biological neural networks; classification; Classification algorithms; Discrete Fourier transforms; Fourier transform; Frequency domain analysis; Microphones; neural network; Neurons; spectrum
• ISSN: 1071-6572; 2153-0742
• DOI: 10.1109/CCST.2010.5678725

In this paper, classification of audio sources is presented to supplement current work on existing system for localization of audio sources. The question of achieving the audio classification lies in the convenient discrimination of the feature vector in the feature vector space. Characteristics based on frequency analysis were chosen and used as feature vector. Artificial neural network was applied in order to classify different audio classes especially from security and military areas, such as different shots and explosions. The information about specific type of a sound can trigger localization process of given audio source. Moreover, it can improve situation when guards get ready for the alert state. This classification method is currently developed as an additional part of the system for audio source hyperbolic localization; the paper also gives some basic structure of that system. Its utilization can be found for additional securing of larger objects like squares or military basis, for instance.