Multimodal Medical Image Fusion Method Based On Clustering And Sparse Representation

Medical image fusion is one of the hot researches in the field of medical imaging and radiology.It has been widely recognized in the medical and engineering fields,and the fusion of two or more medical image information with complementary information has become a hot research direction.This paper is aimed at the improvement of multimodal medical image fusion based on clustering and sparse representation.The main research contents are as follows:Aiming at the problem that the image reconstruction quality is poor due to the large amount of redundant information in the over-complete adaptive dictionary in multi-modal medical image fusion,a multi-modal medical image fusion method based on joint image patches clustering adaptive dictionary learning is proposed.The number of redundant image blocks is reduced by preprocessing.Then,the local gradient information of the image block is extracted as a cluster center,and joint image block clustering is performed.Training is performed on the basis of joint image block clustering to obtain sub-dictionaries,and the sub-dictionaries are merged into an adaptive dictionary.Finally,the sparse representation coefficients are obtained under the action of the adaptive dictionary,and the sparse coefficients are merged to reconstruct the fused image.The result shows that the proposed method makes the fusion image more sharp and contrast.Adaptive learning dictionary atoms are more compact.Aiming at the problem of insufficient detail retention in multimodal medical image fusion based on sparse representation,this paper proposes a multimodal medical image fusion method(CSR-DPC)based on density peak clustering and convolution sparse representation.The medical image is layered to obtain a base layer image and a detail layer image.The detail layer image is fused by convolution sparse representation to obtain the fused detail layer image,and the base layer image is clustered to obtain several clusters,and the adaptive dictionary is trained,and the base layer sparse coefficient is obtained by fusion,and the base layer fused image is reconstructed.Finally,the fused detail layer image is merged with the fused base layer image to reconstruct the final fused image.Experiments show that the proposed method makes the image details clearer,the visual quality is better,and the objective evaluation index is better.A multimodal medical image fusion system based on Python was developed for the two improved methods proposed in this paper.The system is mainly divided into a login interface,a registration interface,and a system main interface.The functions implemented in the main interface of the system are: input source image,selection fusion method,and output fusion result.The fusion method module shows that the two methods proposed in this paper have better fusion effects by comparing different methods.