A Comparative Inspection and Performance Evaluation of Distinct Image Fusion Techniques for Medical Imaging

Fusion in medical imaging refers to the synthesis of medical images obtained from various imaging technologies by utilizing an algorithm to combine the benefits or complementarities of each image to produce a more insightful image. This can greatly improve the accuracy with which diseases are detected and treated for positive health of the patient. In this study, a comparative analysis among widely used image fusion methods in transform domain using a medical image dataset of varying imaging modalities is carried out to determine the finest strategy for future study and give medical researchers new direction. The outcomes of the experiments are evaluated using a range of well-known performance evaluation criteria. Laplacian pyramid provided the best values of \(Q^{{{\text{AB}}/F}}\) in all three datasets, GF-based strategy is most frequently favored by the metrics \(Q_{S}\) , \(Q_{C}\) , \(Q_{Y}\) while \(Q_{P }\) gives the best results for ASR. The hybrid method NSST-MSMG-PCNN gave the second-highest values for all three datasets. Thus, the qualitative and quantitative tests performed on a conventional dataset of multimodal medical images reveal that fusion process produces better quality images and both are necessary for a consistent comparison of fusion methods.