BFO-ICA based multi focus image fusion

Abstract

This paper presents a pixel based multi focus image fusion technique using independent component analysis (ICA) and bacteria foraging optimization (BFO) algorithm. The basic idea here is to obtain the ICA bases from a set of registered images and optimize them using BFO. The novelty in this paper is that BFO-ICA has not been applied to multi-focus image fusion. The images in the ICA domain are fused and the fused image is then reconstructed using inverse transform. The results are compared with FastICA and PSO-ICA. It is observed that optimizing with BFO yield better result.