May 2002

BACK   

One of the greatest challenges for signal and image processing remains the development of efficient representations of signals and images. An efficient representation should permit to study directly the properties of the signal, and understand its structure. Equivalently, the representation should permit to describe the signal with the minimal amount of data.

In this talk we will address the following two problems:

1. the coding of textured images, and,
2. the analysis of functional magnetic resonance images (fMRI) of brain activity.

We will develop methods to convert the original data sets (textured images or fMRI time series) into ``interesting features''. In particular we will provide constructions of collections of basis functions on which the projection of the data will reveal the organization of the sets.

In the first part of the talk, we will demonstrate that there exists many other ways to analyze and represent images that go beyond the standard bi-dimensional discrete cosine transform (used in JPEG) or the discrete wavelet transform (used in JPEG-2000).

In the second part of the talk, we will present a new algorithm to construct ``clustering basis functions''. These basis functions are selected from large libraries of wavelet packets according to their ability to separate the fMRI time-series into ``activated'' and ``non-activated'' clusters.

Publications and softwares can be found on the web-page of the speaker:

• http://ece-www.colorado.edu/~fmeyer

BACK

Last Updated: April 29, 2002