Plugin procedure in segmentation and application to hyperspectral image segmentation

TL;DR

This paper introduces a new plug-in procedure for image segmentation, providing theoretical efficiency conditions, an algorithm, and an application to hyperspectral image segmentation using dimension reduction and spatial regularization techniques.

Contribution

It offers a general framework with efficiency conditions for plug-in segmentation procedures and applies it to hyperspectral images with a novel combination of techniques.

Findings

Algorithm satisfies the efficiency conditions.

Effective segmentation of hyperspectral images demonstrated.

Combines dimension reduction with spatial regularization.

Abstract

In this article we give our contribution to the problem of segmentation with plug-in procedures. We give general sufficient conditions under which plug in procedure are efficient. We also give an algorithm that satisfy these conditions. We give an application of the used algorithm to hyperspectral images segmentation. Hyperspectral images are images that have both spatial and spectral coherence with thousands of spectral bands on each pixel. In the proposed procedure we combine a reduction dimension technique and a spatial regularisation technique. This regularisation is based on the mixlet modelisation of Kolaczyck and Al.