Design and Implementation of Medical Image Fusion Algorithm Based on PET/CT
|Course||Computer Software and Theory|
|Keywords||PET/CT image fusion principal component analysis multiwavelet transform|
In the modern medicine, many medical imaging devices can make lots of image information available, and show the status of body intuitionistily and exactly. We can combine the multimodality medical image information efficiently; display the structures and functions together with changes of the tissue clearly, and provide comple-mentary information by using image fusion. The technology of medical image fusion based on PET/CT is an important method in studying at the metabolism of heart and brain, physical function on the molecular level, which is more and more concernd in the areas such as oncology, cardiovascular diseases and diseases of the nervous system.This paper aims at the study and implemantation on medical image fusion based on PET/CT. In this paper, based on the related technology literatures in the medicine image fusion widely read this thesis has studied the types of the medical image fusion, the image fusion methods, the objective evaluation indexes of the image fusion. Design and implement improved PCA (principal component analysis) image fusion algorithm and multiwavelet transform image fusion algorithm on pixal image fusion. For the imbalance of image, the author develops an improved PCA medical image fusion algorithm based on PCA image fusion by using image block preprocessing. In addittion, as the result of advantages of multiwavelet in image analysis and processing beyond single wavelet, design and implemant the multiwavelet transform medical image fusion algorithm based on the further research of discrete multiwavelet transform, the nature and coefficients’fusion strategies of multiwavelet.The experiments of the fusion algorithms in this thesis have been carried on under the Microsoft Visual C++6.0 environment, and the objective evaluation’s result of the fusion has been given. The result of the experiments shows that the both methods can fuse details of PET/CT images successfully by different ways and get more satisfactory result. To some extent, these two methods makes function on clinical medicine.