Blind Separation of Permuted Alias Image Base on Four-phase-difference and Differential Evolution

Abstract

Blind separation of permuted alias image was a new type of single channel blind separation, which was fundamentally different from traditional single channel blind separation in theory and method. In this paper, a blind separation algorithm based on four-phase-difference and differential evolution was proposed for a type of permuted alias image with blur difference. The model of permuted alias image was briefly introduced. Space domain was selected as characteristic domain after analyzing various detection methods of blur characteristic. First, four-phase-difference versions of permuted alias image are computed and the fitness function was first evaluated by calculating sum of square errors of binary version and its version with a threshold vector. The differential evolution was then performed to obtain the optimal threshold vector for threshold the differential image into binary one. The permuting image could be separated by permuted alias image multiplied with the binary image. Experimental results showed that the proposed approach could effectively separate the permuting image from the permuted alias image without respect to the location, size, number and blurry types of the permuting image.

Keywords

• Blind separation
• Permuted alias image
• Four-phase-difference
• Differential evolution.