Chapter 14 - Multimodality medical image fusion in shearlet domain

Abstract

For various medical applications, the fusion of important imaging information has emerged as a critical issue. A new multimodality medical image fusion technology, referred to as the shearlet domain, is proposed in this study. The proposed technique decomposes input images using nonsubsampled shearlet transform (NSST) to extract low- and high-frequency components. A unique approach, which includes bilateral filter processing and local energy-based fusion, is applied to acquire crisp and smooth features in low-frequency components. An SML algorithm (sum modified Laplacian) combines the high-frequency coefficients to provide image fusion for high-frequency components. This study explores the experimentation results and the analysis of many medical modalities on the multimodal medical image dataset. The proposed methodology outperforms the cutting-edge fusion algorithms that deal with edge preservation in objective and subjective assessments.