An extraction method for digital camouflage texture based on human visual perception and isoperimetric theory

Abstract

To deal with the problem of lack of subjective visual perception in texture features exacting of digital camouflage, a new extraction arithmetic based on human visual perception and isoperimetric theory is proposed in this paper. The method firstly constructs edge weight function according to human visual perception, then selects isoperimetric ratio as a as a criterion to determine the optimal threshold from the candidates, finally utilizes an iterative scheme to select multiple thresholds in order to segment image into multi-regions. The experimental results of segmentation show that our method is more effective than current threshold methods in segmentation quality.