Overlapped soundtracks segmentation : a singular spectrum analysis and random forests approach

Mohammed, DY and Li, FF 2017, Overlapped soundtracks segmentation : a singular spectrum analysis and random forests approach , in: The 2nd International Conference on Knowledge Engineering and Applications, 21-23 October 2017, Imperial College London, London, UK.

[img] PDF - Accepted Version
Restricted to Repository staff only

Download (412kB) | Request a copy
Official URL: http://www.ickea.org/

Abstract

In the field of audio classification, audio signals may be broadly divided into three classes: speech, music and events. Most studies, however, neglect that audio data from the real world can have any combination of these classes simultaneously. This study is to adapt existing methods by reducing the degree of overlap between different classes of audio content, in order to mitigate classification difficulties and improve the performance of automatic classification. In this work, singular spectrum analysis method serves to mitigate the overlapping ratio between speech and music in the mixed soundtracks by generating two new soundtracks with a lower level of overlapping. Next, feature space is calculated for the output audio streams, and these are classified using random forests into either speech or music. The classification performance of overlapped soundtracks is effectively improved and singular spectrum analysis has been found to be an efficient way to discriminate speech/music in mixed soundtracks.

Item Type: Conference or Workshop Item (Paper)
Schools: Schools > School of Computing, Science and Engineering
Journal or Publication Title: The 2nd International Conference on Knowledge Engineering and Applications
Depositing User: Duraid Yehya Mohammed
Date Deposited: 28 Jun 2017 14:38
Last Modified: 09 Aug 2017 01:33
URI: http://usir.salford.ac.uk/id/eprint/42801

Actions (login required)

Edit record (repository staff only) Edit record (repository staff only)

Downloads

Downloads per month over past year