Multispectral Image Recognition Research Based on Biomimetic Pattern Recognition

Abstract

Studies on learning problems from geometry perspective have attracted an ever increasing attention in machine learning, leaded by Biomimetic Pattern Recognition on information geometry. Biomimetic Pattern Recognition is a new model of Pattern Recognition based on “matter cognition” instead of “matter classification”. This new model is much closer to the function of human being, than traditional statistical Pattern Recognition using “optimal separating” as its main principle.This paper proposes a high-dimensional descriptive way about multispectral image. Geometrical properties of high-dimensional structures underlying a set of samples are learned via successive projections from the higher dimension to the lower dimension until Clifford space, under guidance of the established properties and theorems in high-dimensional descriptive way. Specifically, we introduce a principle of homology continuity based on Clifford Algebra and provides a geometrical learning algorithm for specifying Clifford Algebra shapes, which is then applied to biomimetic pattern recognition. Experimental results are presented to show the efficiency of our theory.