The development of robust algorithms for the recognition and classification of sensory data is one of the central topics in the area of intelligent systems and computational vision research. In order to build better intelligent systems capable of processing environmental data accurately, current research is focusing on algorithms which try to model the types of processing that occur naturally in the human brain. In the domain of computer vision, these approaches to classification are being applied to areas such as facial recognition, object detection, motion tracking, and others. This project investigates the extension of these types of perceptual classification techniques to the realm of acoustic data. As part of this effort, an algorithm for audio fingerprinting using principal component analysis for feature extraction and classification was developed and tested. The results of these experiments demonstrate the feasibility of such a system, and suggestions for future implementation enhancements are examined and proposed.
Library of Congress Subject Headings
Sound--Classification; Principal components analysis; Computer vision
Burka, Zak, "Perceptual audio classification using principal component analysis" (2010). Thesis. Rochester Institute of Technology. Accessed from
RIT – Main Campus