Edge-aware multisensor brain image fusion via guided filtering in Laplacian domain

Medical image fusion is crucial in clinical diagnosis since it enhances diagnostic accuracy by integrating complementary data from many modalities, including Computed Tomography (CT), Magnetic Resonance Imaging (MRI) and Positron Emission Tomography (PET). However, deformities like noise, spatial distortion and the loss of feature fine details during fusion remain, necessitating robust approaches. This work presents a new edge-aware multi-sensor brain image fusion technique using Laplacian energy maps for edge preservation, guided filtering for image decomposition, and adaptive fusion algorithms for integration of base and detail layers. Guided filtering divides source images into base and detail layers such that structure is preserved while noise is reduced. Detail layers are combined using Laplacian energy maps to highlight textures and edges whereas base layers are fused by weighted averaging to balance intensity and contrast. The proposed method was tested on multi-modal datasets and demonstrated significant improvements in fusion quality. These improvements indicate better edge preservation, structure preservation, and diagnostic accuracy, offering a clinically viable solution for tumor detection, surgical planning, and neurological assessments.