A multi‐modality paradigm for CT and MRI fusion with applications of quantum image processing. (23rd September 2021)

• Record Type: Journal Article
• Authors: Dogra, Ayush; Ahuja, Chirag Kamal; Kumar, Sanjeev
• Other Names: Wright Steven A. guestEditor.; Solak Serdar guestEditor.; Kilimci Zeynep Hilal guestEditor.; Eken Süleyman guestEditor.; Fernandes Steven guestEditor.; Zhang Yu‐Dong guestEditor.; Tavares João Manuel R.S. guestEditor.
• Is Part Of: Concurrency and computation. Volume 34:Number 20(2022)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2021-09-23
• Subjects: CT -- medical image fusion -- medical image processing -- MRI -- quantum image processing -- spatial filtering; Parallel processing (Electronic computers) -- Periodicals; Parallel computers -- Periodicals; 004.35
• DOI: 10.1002/cpe.6610
• Languages: English
• ISSNs: 1532-0626

Abstract

Abstract: As pixel resolution continues to increase in images acquired from different sensors, the measurement of the correlation between pixels and their properties requires numerous processing steps over multiple iterations. Quantum image processing (QIP) is a technology capable of processing exponentially large volumes of data within polynomial time. One such application is image fusion. In the field of radiography, computerized tomography (CT) images provide the structural information of bone tissues, and magnetic resonance MR images can provide the visualization of gray‐scale anatomical structures of soft tissues. With the help of CT and MR fusion, a composite image can be obtained containing information on both hard and soft tissues. Clinicians often have to switch between these modalities to trace various patterns for effective staging and diagnosis of various oncological diseases. In this study, a multilevel filtering‐based image fusion algorithm that efficiently fuses information from CT and MR images into a single image has been proposed. First, source images are processed with a rolling guidance filter, and the detail layer is calculated. The base layer is then computed using guided filtering to ensure a high level of edge preservation. Furthermore, to minimize noise and artifacts, the detail layers are fused using the Karhunen–Loeve transform, and the base layers are combined using the weighted superimposition principle. Experimental results demonstrate that the proposed methodology is subjectively more efficient than the existing state‐of‐the‐art methods.