Medical Image Fusion Based On Two-scale Decomposition And Saliency Detection

Various medical images in multi-imaging modes have different functional characteristics,and single-modal medical images usually cannot provide enough information.Therefore,fusion of medical images with different characteristics to assist diagnosis has become a research hotspot,such as the image fusion of magnetic resonance imaging(MRI)images and computed tomography(CT)images,the image fusion of MRI images and single single photon emission computed tomography(SPECT)images,and the image fusion of MRI images and positron emission computed tomography(PET)images.However,due to its fusion algorithm setting,it may cause some visual problems in the fusion image to affect the doctor’s diagnosis.For the problems caused by different types of image fusion algorithms,this paper designs two algorithms to fuse multi-modal medical images.The result images of the current image fusion algorithms for MRI images and CT images have problems with texture loss,low contrast,and pseudo-edges.In order to overcome the shortcomings of the fusion algorithms,this paper proposes a medical image fusion method based on bilateral texture filtering(BTF),structure similarity detection(SSD)and saliency detection.First,BTF is used to decompose the source image into a base layer and a detail layer.For the base layer fusion,a fusion rule based on contrast saliency detection is designed to retain structural information.For the detail layer fusion,the method of SSD is proposed,and the structurally similar parts and structurally dissimilar parts are obtained respectively.The former uses an improved absolute maximum strategy(IAMV)for fusion,and the latter uses the saliency estimate(SEM)algorithm for fusion.Finally,the fused image is obtained through reconstruction.Through subjective and objective comparison and evaluation of the algorithm with the other 12 algorithms,the effectiveness of the algorithm is verified.In the current image fusion of MRI images and PET images,as well as image fusion of MRI images and SPECT,the fused images have disadvantages such as texture loss,artifacts and color distortion of the fusion images.In order to avoid the above-mentioned problems with fused images,this paper proposes a medical image fusion algorithm based on relative total variation(RTV)decomposition and saliency detection.Firstly,RTV is used to decompose the MRI image into the base layer and the detail layer.Secondly,the base layer and functional image are transformed into YUV coding.Then the saliency detection of the two Y components is performed,the fusion base layer is obtained according to the saliency map,and the fusion Y component is obtained by combining the base layer and the MRI texture layer.Finally,the fusion Y component and the color components U and V of the functional image are subjected to YUV inverse transformation to obtain the final RGB fusion image.The advantage of the algorithm is that RTV decomposition extracts the detail layer,and can well retain the texture of the fused image.At the same time,YUV coding is used to avoid the color distortion problem of the traditional RGB coding,so that the color information is retained in the fused image.This paper uses a combination of subjective and objective methods to compare the algorithm with other 6 algorithms.The experimental results show that the algorithm is superior to the comparison algorithm in visual results and objective analysis.