We present a method to isolate and differentiate objects of different contrasts, sizes and structures in complex, multi-scaled sensor imagery. The analysis is based on scale and spatial information of spectral TV bands, referred to as spectral local scale signatures. We use dimensionality reduction to extract highly contrasted objects in thermal and medical images for image fusion; as well as groups of objects of similar features in images of repetitive structures for image manipulation.

This presentation is part of Minisymposium “MS32 - Nonlinear Spectral Theory and Applications (part 1)”
organized by: Aujol Jean-Francois (University of Bordeaux) , Gilboa Guy (Electrical Engineering Department, Technion) .

Authors:

Hait Ester (Technion)

Keywords: