Manipulation of compressed data using MPEG-7 low level audio descriptors

Abstract

This paper analyses the consistency of a set of MPEG-7 low level audio descriptors when the input audio stream has previously been compressed with a lossy compression algorithm. The analysis results show that lossy compression has a detrimental effect on the integrity of practical search and retrieval schemes that utilize the low level audio descriptors. Methods are then proposed to reduce the detrimental effects of compression in searching schemes. These proposed methods include improved searches, switched adaptive scalar and vector prediction, and other prediction schemes based on machine learning principles. Of the proposed schemes the results indicate that searching which incorporates previous and future frames combined with machine learning based prediction best nullifies the effects of compression. However, future scope is identified to further improve the reliability of the MPEG-7 audio descriptors in practical search environments.