Geometric-Based Segmentation of Polarization-Encoded Images

Abstract

In the framework of Mueller parameters imaging, polarization-encoded images have sixteen channels. The relevancy of such multidimensional structure comes from the set of physical information they carry about the local nature of the target. The admissibility constraints imposed on these images make awkward their analysis and processing and prevent to explore their richness. This induces the need for a proper tool that allows the analysis and processing of polarization-encoded images. In this paper we address a new method to segment Mueller imaging and use the geometric algebra to represent the polarization formalism and segment polarization-encoded images while respecting their physical meaning. The segmentation task is based on the fuzzy K-mean algorithm.