Multi-Focus Image Fusion Method Using Higher Order Singular Value Decomposition and Fuzzy Reasoning

Abstract

Higher order singular value decomposition (HOSVD) is an efficient data-driven decomposition technique, and shows the salient ability in the representation of high-dimensional data and feature extraction. In addition, fuzzy reasoning can solve the uncertainties of the source images' contributions to the fused image and is easy to apply. Motivated by the advantages mentioned above, a new HOSVD and fuzzy reasoning-based multi-focus image fusion method is proposed. Firstly, sub-tensor is constructed by two image patches separately from the two multi-focus images and HOSVD is employed to obtain the decomposition coefficients of image patches. Secondly, the weighted average fusion rule based on fuzzy reasoning is proposed for fusing the decomposition coefficients, and fuzzy reasoning rule is designed based on average energy, regional energy and match degree. Finally, the fused image is achieved by the inversing HOSVD. Experimental results indicate that the proposed method performs better than other methods both visually and quantitatively.