Brain Medical Image Registration And Fusion Algorithm Research

In recent years,modern medical imaging equipment has been developed rapidly with the progress of science and technology.Different medical imaging devices can produce different modes of medical images,and different modes of medical images can reflect the specific information of different organizational structures of the human body.Through information integration of different multimodal images,doctors can observe more useful information on the same image,so as to be able to better detect and diagnose diseases.Therefore,medical images are very important Image registration and fusion technology came into being.In this paper,the algorithm of medical image registration and fusion is studied.(1)Aiming at the problems of the existing medical image registration algorithms,such as low registration accuracy,low image matching and image offset after registration,this paper proposes an automatic medical image registration method based on region of interest(ROI).In this method,the common sub regions of interest in the two images(target image and floating image)are automatically detected and registered,and then the conversion parameters obtained from the sub region registration are used.Finally,the conversion parameters are applied to the overall image registration to achieve the final registration of the source image.The experimental results show that the proposed image registration algorithm has good performance,high accuracy,small offset,better matching the target image and floating image,and the registration time is significantly better than other image registration algorithms.(2)Aiming at the problem of poor fusion effect caused by the loss of detail features in the fusion process of existing medical image fusion algorithms,this paper proposes a parameter adaptive pulse coupled neural network,A multimodal medical image fusion method based on PAPCNN and CSR(convolutional sparse representation)is proposed.Firstly,the medical source image to be fused is transformed by non-subsampled shearlet transform transform(NSST)is decomposed into high and low frequency components,and then the high frequency components decomposed from the source image are fused by using the improved papcnn;then the low frequency components of the source image are fused by using CSR;finally,the fused high and low frequency components are reconstructed into the fused image by using the NSST inverse transform.The experimental results show that the proposed medical image fusion algorithm can better fuse the detail features and edge texture information of the source image than other contrast algorithms,and also has better performance in the objective evaluation index.