Research On Medical Image Fusion Algorithm Based On Multi-scale Decomposition And PCNN

Medical images are important tools to assist doctors in daily diagnosis.Medical images with different modes are used to monitor different information of the human body,such as bone,soft tissue or human metabolism.In actual clinical diagnosis,doctors often need to combine all kinds of information of the human body to determine the lesion,while a single mode of medical image can only describe a certain aspect of the human body information,which affects the efficiency and accuracy of medical diagnosis.In order to overcome this defect,multimodal image fusion technology has been applied in the field of medical image fusion.By combining different modes of medical images,important information from different modes of medical images can be concentrated in one image,making medical diagnosis more accurate and convenient.In this dissertation,We studied the image fusion technology in existence and then combined the multi-scale decomposition technology with Pulse Coupled Neural Network(PCNN)which was proposed to simulate the vision principle of mammals,and proposed two image fusion algorithms based on the combination of multi-scale decomposition and PCNN.The main work is as follows:(1)Aiming at the problem that the PCNN model affects the fusion result due to complex parameter settings during the fusion process,in order to further improve the quality of the fusion image,A medical image fusion algorithm based on NSST and optimized PCNN by QPSO is proposed.First,the source image after NSST decomposition can get a low frequency subband and a series of high frequency subbands,and calculate an image of a region of low frequency subband energy to keep the contour information,using QPSO algorithm to optimize the PCNN model fusion high-frequency subbands in order to keep the details of image,in the end,the final fused image is obtained by NSST reverse transform.The experimental results show that the method can effectively retain the image details in the subjective vision,and has a significant improvement in the objective indicators.(2)Aiming at the problem that the traditional multi-scale decomposition technology could not protect the edge of the image during the decomposition process,which led to the blurring and artifact of the fused image at the edge,A medical image fusion algorithm based on multiscale edge-preserving decomposition and PCNN is proposed.For the multi-scale image decomposition,a new decomposition method based on the weighted least squares filter,combined with Gaussian filter and non-subsampled direction filter bank is proposed,the source images are decomposed into a low frequency layer and a series of high frequency layers,using the PCNN with a biological background as a fusion rule,and in view of the high frequency and low frequency have different information,improved spatial frequency and regional energy are calculated respectively as the external incentive of PCNN.Experimental results show that this method can effectively retain the edge structure information and enhance the image details.(3)A multimodal medical image fusion system is designed to apply the research into practice.The system is implemented using Graphical User Interface in Matlab and sets different fusion modules according to different types of medical image fusion,selects different fusion methods in the image fusion module for fusion,and the objective evaluation index is used to evaluate the fusion images obtained by different fusion method to select higher quality fusion images.