Proposal for preservation criteria to rare event. Application on multispectral/hyperspectral images

Abstract

Rare events can potentially occur in many applications, particularly in hyper spectral image analysis. When manifested as opportunities to be exploited, such events become of paramount significance. Due to their sporadic nature, the information-bearing spectral vectors associated with rare events often lie in a large set of irrelevant signals and are not easily accessible This paper provides a statistical frame of a local criterion to measure the rate of the conservation of these rare events during of the dimension reduction. The core component of this framework is a sampling procedure that adaptively and quickly focusses the information-gathering resources on the segments of the data set that bear the information pertinent to the rare events. A particular accent is put on the local vectors in every spectral band with the aim of measuring the rate of the conservation of the rare vectors.