Multifocus image fusion based on discrete fuzzy transform

Abstract

Multifocus image fusion has emerged as a challenging research area in the field of machine vision and image processing. Processing is required of images obtained through various image capturing devices. The optical lenses that are widely used in image capturing devices have limited ‘depth-of-focus’ and therefore only the objects that lie within a particular depth remain ‘in-focus’ while all the other objects are ‘out-of-focus’. Thus, with a view to obtain an image with all the objects ‘in-focus’, multifocus image fusion algorithm based on discrete fuzzy transform (D-FTR) is proposed. The proposed fusion algorithm uses two different rules: weighted average rule and select maxima rule to fuse images in D-FTR domain. The final fusion rule is determined using a matching coefficient and the square root of variance. The performance of the proposed fusion algorithm is evaluated by performing several experiments and comparing their results with the results of the recent state-of-art image fusion algorithms. Results demonstrate that the proposed algorithm is not only effective in producing the fused images that are best focussed everywhere but also performs better than the recent state-of-art fusion algorithms.