Multimodal MRI and PET /SPECT Brain Image Integration for improved clinical diagnosis of PACS in Health care

• Saravanan Vijayan,
• Malarvizhi Subramani,
• Senthil Kumar Aiyappan,

Vol. 19, No. 7, July 31, 2025
10.3837/tiis.2025.07.011, Download Paper (Free):

• Computer-Aided Diagnosis
• MRI-PET fusion
• Medical Image Processing
• multiscale transform
• PACS

Abstract

Multimodal medical image fusion is an essential technological advancement in medical diagnosis that integrates data from several image modalities, giving clinicians an improved visual perception for earlier illness identification. This work aims to improve the computer-aided diagnosis CAD system using image processing. We propose a two-level decomposition method utilizing Discrete Wavelet Transform (DWT) and Non-subsampled Contourlet Transform (NSCT) for MRI-SPECT/PET image fusion for improved visual quality. In the subsequent stage, the phase components PC of the images were computed and combined with the first stage output. A final single fused image with good clarity and sharpness is obtained by applying inverse transforms, which helps doctors for quicker identification and diagnosis of brain tumour diseases. This consecutive stage of decomposition along with phase congruency provides an in-depth analysis of frequency components of source images with edge retention which helped in extracting the soft cell portions and tumour-affected portions effectively with less computational complexity than the other related articles including the recent deep learning techniques. Images of brain diseases in the Whole Brain Atlas dataset and real-time test cases of MRI and SPECT images were taken for performance study and got subjective assessment verified with the radiologist. Qualitative and quantitative metrics were computed with similar works and our proposal provides an improvement of visual quality parameters. The experimental analysis proved that the effectiveness of our fusion approach outperformed well compared to existing methodologies. Our algorithm can be deployed with the existing PACS.

Statistics

Cite this article