| At present,in medical images,different modalities of medical images can be obtained through different sensors.These images are all auxiliary means for doctors to treat diseases,but different modalities of medical image imaging focus on different aspects.For example,CT images in anatomical images reflect the structural information of organs and tissues,and MRI images reflect the structural information of organs and tissues.It is soft tissue information,while functional images such as PET and SPECT images capture projection data at low resolution and reconstruct tomographic images to provide tissue,blood and metabolic information of related organs.The purpose of medical image fusion is to fuse medical images with different imaging focuses into one image,which can integrate the important information of the original image,simplify the doctor’s diagnosis process,and optimize the treatment.Based on this research purpose,this paper studies the multimodal medical image fusion algorithm.The main work includes the following two aspects:(1)Aiming at the problems of low contrast and insufficient retention of important details caused by medical image fusion,a Non-Subsampled Shearlet Transform(NSST)and Pulse-coupled Neural Network(PCNN)combined multimodal medical image fusion method.First,the source image is decomposed by non-subsampling shear wave transform to obtain its high frequency subband and low frequency subband;Secondly,for the high frequency subband image,the image detail feature information is used as the external excitation condition to stimulate the pulse coupled neural network.In order to achieve fusion;for lowfrequency subband images,a weighted average strategy based on Visual Saliency Mapping(VSM)is used for fusion;Finally,the final multimodal medical image fusion is obtained using inverse shearlet transform.The experimental results show that compared with the existing five representative medical image fusion methods,it has certain advantages in subjective and objective evaluation.(2)Aiming at the problem that the current medical image fusion algorithms mostly use multi-scale decomposition,the edge feature information of medical images is not effectively preserved,resulting in halo at the edge of the image,a multi-level edge decomposition and improved local gradient energy are proposed.fusion algorithm.Firstly,the source image is decomposed under the proposed multi-level edge-preserving decomposition framework theory,and the high and low frequency layer images of the image are obtained;Secondly,the low frequency layer combining the local gradient energy operator and the weighted Laplacian operator is given.The fusion rule uses the Multi-Scale Morphological Gradient(MSMG)of the image as the link strength of the PCNN network to influence the high-frequency layer of the PCNN fusion to realize the fusion of medical images.Finally,through comparative experiments,this method can more highlight the edge contour information of the image and retain the relevant details of the image,which has certain advantages compared with the current seven representative image fusion algorithms. |
