Some Trends in Application of Geometric Approaches in Multimodal Medical Image Fusion

The multimodal medical fusion combines data from different medical modalities, and thus can improve clinical information from medical images for a better understanding of the content of the image and to obtain a more precise diagnosis. The multimodal registration task is not easy, since images obtained from different modalities may have exceptional differences. This type of regeneration produces images that combine physiological and anatomical information. The aim of this paper is to present some trends in application of geometric approaches in fusion of multimodal medical images. For the efficient extraction of geometric characteristics, some trends are related to the use of structure tensors in combination with other transformations. The problem is considered in two main directions, namely registration and sparse representation and their role for obtaining better results in multimodal medical image fusion.