I understand you want to perform Image decomposition on an image to split it into base and detail layer and at the same time preserve more edges in the process.
One commonly used technique is the application of multi-scale decomposition methods, particularly those based on the concept of multiscale transforms. These methods allow you to extract both the low-frequency components (base layer) and high-frequency components (detail layer) of the image.
Start by loading and preprocessing both images to ensure they have consistent dimensions and intensity ranges. Perform wavelet decomposition on both images using the wavedec2 function, specifying the desired wavelet family and decomposition level.
Combine the wavelet coefficients from the two images, typically using weighted averaging to control the contribution of each modality. Reconstruct the fused image from the combined coefficients using waverec2.
Refer to this documentation for more information:
