Edge-preserve Filter Image Enhancement With Application To Multi-modal Medical Image Fusion

Multimodal medical image fusion is the process of merging multiple images from multiple imaging modalities about the same target into a single image,and to keep the important information of each image.Currently,multimodal medical image can be divided into two categories: the anatomy of the medical image and the functional of medical images.Anatomy of the medical images(such as CT,MRI)has high resolution,and can provide the structural information of human organs clearly,but can't reflect the function of the organ.Function of medical images(such as PET and SPECT)could reflect the metabolic activity of the human body tissues and organs,but due to its low resolution,image is not clear,it is not easy to observe organs structure and lesion location information.Multimodal medical image fusion technology can effectively combine the advantages of different modal medical images and reduce redundant information,and help medical staff to observe the location of lesions more directly and clearly,so as to make a reasonable diagnosis.Multimodal medical image fusion method has achieved some great results,but still exists some shortcomings.The fusion method which is based on principal component analysis transform is simple and practicable but the fusion results produce color distortion;The fusion results of fusion method based on color space has a low contrast and always based on tower type transform fusion method which can not capture the direction information;Wavelet transform can't show anisotropic characteristics(line,outline,etc.)well;Although some other image fusion methods are successful applied to multimodal medical image fusion,but they all have their own shortcomings.Aim at the shortcomings of the existing image fusion method in the process of multimodal medical images fusion,in this thesis,on the basis of edge preserving filter,my research work is as follows:1.This thesis proposes a medical image fusion framework which based on the edge preserve filters and Multi-scale transform fusion methods.The framework utilizes edge preserve filter enhance edge and detail information of input images,and then adopt multiscale transform fusion method to combine enhanced images.Compared with the traditional filter,edge filter as a tool for enhancing image preprocessing,improves the image edge and significant characteristic information and control the impact of noise at the same time.Experimental results show that the fused image by porposed fusion framework is in high contrast and clear edge.2.In this thesis,a multi modality medical image fusion method based on local Laplasse Pyramid decomposition is proposed.The proposed method firstly adopt local Laplacian Pyramid decompose input images.Then,adoptive cloud model is used to fuse lowfrequency images and salience match measure is used to fuse high-frequency images.Finally,inverse local Laplacian Pyramid transform is used to image reconstruction.In the case of fusing MRI-PET and MRI-SPECT,the proposed method with good noise robustness and the saliency characteristic is more obvious.