Research On Multimodal Medical Image Fusion Method Based On Multiresolution Analysis

At present,with the rapid development of medical technology,people have higher and higher requirements for clinical diagnosis technology.Medical image fusion technology can effectively overcome the limitations and differences of a single sensor image,acquire more comprehensive and accurate image information,improve the clarity of the image,so as to further analyze and process the image,which has a wide range of research significance and practical value for clinical medical diagnosis.At the initial stage,image fusion technology is mainly used in the field of military security,which can increase Because of its advantages,image fusion has been widely used in medical imaging in recent years.Medical image is complex and diverse.In this paper,according to the characteristics of sub image coefficients under different scales and directions of multi-scale transformation,the medical image fusion algorithm under the framework of multi-scale decomposition is studied,mainly from the following aspects:(1)In order to solve the problem that the detail texture of the fused medical image is not clear enough,this paper proposes a new medical image fusion algorithm based on non subsampling shear wave transform,which can fuse the multimodal medical image,enhance the ability of detail structure extraction,improve the image fusion quality,and provide the basis for medical diagnosis.Firstly,the registered source image is decomposed by NSST to obtain low frequency subband and a series of high frequency subband;secondly,for low frequency subband coefficients,a fusion strategy is proposed to select the subband by using the composite value of local average energy and local standard deviation;for high frequency subband coefficients,the improved Laplace energy sum method is used for fusion;then,the fused low The high frequency subband is inverted by NSST to get the fused image.(2)At present,most of the traditional multi-modal medical image fusion technology can not achieve energy conservation and detail extraction at the same time,which is easy to lead to information shortage and detail blur,etc.in this paper,a medical image fusion algorithm based on parameter adaptive pulse coupled neural network is proposed.Firstly,four different types of registered source images are extracted from the standard database for non subsampling shear wave decomposition to obtain multiscale and multidirectional low-frequency subband coefficients and a series of high-frequency subband coefficients;secondly,the high-frequency subband is fused by parameter adaptive PCNN,and the enhanced high-frequency subband absolute value is used as the feedback input to adaptively adjust all PCNN parameters.The low-frequency subband uses the improved average gradient and spatial frequency to fuse the image,which can save energy and extract details at the same time.Finally,the fused medical image is obtained by NSST inverse transformation.(3)In order to test the effectiveness and generality of the image fusion algorithm,this paper conducts a large number of fusion experiments on gray image and color image respectively,and compares the fusion results with the fusion results generated by typical different fusion algorithms.It is not difficult to find that the algorithm in this paper can play a good complementary role for medical images between different modes by naked eye observation,making the image details more Rich,more clear and complete information display for edge position,texture contour and other information,good visual effect,which can play an important role in the clinical diagnosis of doctors;quantitative analysis of image quality from an objective point of view,it is not difficult to see that the algorithm proposed in this paper has advantages in information entropy,edge strength,standard deviation and other index values,and the index values are generally good,which is shown in this table After the fusion,the image information integrity and structure are better,the detail information is rich and comprehensive,and the detail texture is clear.