Wilson, R.G.,
Calway, A.D.,
Pearson, E.R.S.
and Davies, A.R.
(1992)
An introduction to the multiresolution Fourier transform and its applications.
Technical Report.
Department of Computer Science, Coventry, UK.
Abstract
The Multiresolution Fourier Transform (MFT) is a new tool for signal analysis which is designed with the express intent of providing a local signal representation adapted to the problem of segmenting signals into s set of meaningful primitive features. It is shown that this requires a signal representation that is richer than conventional Short Time Fourier Transforms (STFT) or indeed Wavelet Transformations (WT), but shares with them the property of giving a phase-space description of the signal. These considerations lead to the new transform, which can be seen as a marriage of the STFT and WT which overcomes the limitations of either as a signal representation for segmentation. Some of the elementary properties of the MFT are then discussed and its implementation for discrete signals is considered. The report is concluded with a summary of results obtained with the MFT on audio and image signal segmentation.
Actions (login required)
 |
View Item |
![[img]](http://eprints.dcs.warwick.ac.uk/style/images/fileicons/application_pdf.png)
