Texture Based Multifocus Image Fusion Using Interval Type 2 Fuzzy Logic

Abstract

Multifocus image fusion is a process of fusing two or more images where region of focus in each image is different.   The objective is to obtain one image which contains the clear regions or in-focus regions of each image. Extracting the focused region in each image is a challenging task. Various techniques are available in literature to perform this task. Texture is one such feature which acts as a discriminating factor between focused and out-of-focus regions. Texture based image fusion has been used in our approach in combination with interval type 2 fuzzy logic and discrete wavelet transforms. Performance metrics obtained using this approach are better compared to other existing techniques. Gray Level Cooccurence Matrix (GLCM) method is used to extract the texture. Type 2 Sugeno fuzzy logic is used to combine the images. The fused image is compared with the reference image when it is available. It is also compared with the original images and performance metrics are computed and presented in this paper.

Keywords: Discrete Wavelet Transform, Gray Level Cooccurence Matrix, Image Fusion, Multifocus Image, Type 2 Fuzzy Logic, Mamdani FLS, Sugeno FLS