Image Fusion using Joint Sparse Representations and Coupled Dictionary Learning

Abstract

We address the multi-focus image fusion problem, where multiple images captured with different focal settings are to be fused into an all-in-focus image of higher quality. Algorithms for this problem necessarily admit the source image characteristics along with focused and blurred features. However, most sparsity-based approaches use a single dictionary in focused feature space to describe multifocus images, and ignore the representations in blurred feature space. We propose a multi-focus image fusion approach based on sparse representation using a coupled dictionary. It exploits the observations that the patches from a given training set can be sparsely represented by a couple of overcomplete dictionaries related to the focused and blurred categories of images and that a sparse approximation based on such coupled dictionary leads to a more flexible and therefore better fusion strategy than the one based on just selecting the sparsest representation in the original image estimate. In addition, to improve the fusion performance, we employ a coupled dictionary learning approach that enforces pairwise correlation between atoms of dictionaries learned to represent the focused and blurred feature spaces. We also discuss the advantages of the fusion approach based on coupled dictionary learning, and present efficient algorithms for fusion based on coupled dictionary learning. Extensive experimental comparisons with state-of-the-art multi-focus image fusion algorithms validate the effectiveness of the proposed approach. Index TermsSparse representations, coupled dictionary learning, image fusion, multi-focus image.The authors are with Aalto University, Dept. Signal Processing and Acoustics, FI-00076, AALTO, Finland. E-mails: farshad.ghorbaniveshki@aalto.fi and svor@ieee.org

Efficient Online Convolutional Dictionary Learning Using Approximate Sparse Components

“…where {s p 1 } P p=1 and {s p 2 } P p=1 are the multi-measure signals (e.g., vectorized overlapping patches extracted from multi-measure images), D 1 and D 2 are the coupled dictionaries, and θ is the maximum number of nonzero entries in common sparse representations {x p } P p=1 . Coupled dictionary learning has been used in various applications, including image super-resolution [2,68,69], image reconstruction [70,71], change detection [72], and image fusion [73,74].…”

Colorization of Multispectral Image Fusion using Convolutional Neural Network approach

“…F.G. Veshki and et al [22] represented generated input images, the JSR method is used to build global and local saliency maps. The researchers then suggested a saliency detection method that merges global and local saliency maps into a single saliency map.…”

Coupled Feature Learning for Multimodal Medical Image Fusion

“…CDL is particularly suitable for tackling problems that involve image reconstruction in different feature spaces. The standard problem (1) has been successfully employed in numerous image processing applications, such as image superresolution [18], single-modal image fusion [19], or photo and sketch mapping [20]. In this section, the objective is to learn the correlated features in two input multi-modal images.…”