Joint medical image fusion, denoising and enhancement via discriminative low-rank sparse dictionaries learning. (July 2018)

• Record Type: Journal Article
• Authors: Li, Huafeng; He, Xiaoge; Tao, Dapeng; Tang, Yuanyan; Wang, Ruxin
• Is Part Of: Pattern recognition. Volume 79(2018:Jul.)
• Page Start: 130
• Page End: 146
• Publication Date: 2018-07
• Subjects: Medical image fusion -- Denoising -- Dictionary learning -- Low-rank decomposition -- Sparse representation; Pattern perception -- Periodicals; Perception des structures -- Périodiques; Patroonherkenning; 006.4
• DOI: 10.1016/j.patcog.2018.02.005
• Languages: English
• ISSNs: 0031-3203

Abstract

Highlights: We develop a method for simultaneous medical image fusion, denoising and enhancement. We design a novel discriminative dictionary learning model for image decomposition. We develop a model to decompose a medical image into low-rank and sparse components. Abstract: Medical image fusion is important in image-guided medical diagnostics, treatment, and other computer vision tasks. However, most current approaches assume that the source images are noise-free, which is not usually the case in practice. The performance of traditional fusion methods decreases significantly when images are corrupted with noise. It is therefore necessary to develop a fusion method that accurately preserves detailed information even when images are corrupted. However, suppressing noise and enhancing textural details are difficult to achieve simultaneously. In this paper, we develop a novel medical image fusion, denoising, and enhancement method based on low-rank sparse component decomposition and dictionary learning. Specifically, to improve the discriminative ability of the learned dictionaries, we incorporate low-rank and sparse regularization terms into the dictionary learning model. Furthermore, in the image decomposition model, we impose a weighted nuclear norm and sparse constraint on the sparse component to remove noise and preserve textural details. Finally, the fused result is constructed by combining the fused low-rank and sparse components of the source images. Experimental results demonstrate that the proposed method consistently outperforms existing state-of-the-art methods in terms of both visual and quantitative evaluations.