The Application Of ICA In PET Image Denoising

Independent Component Analysis (ICA) is a kind of powerful method for Blind Signal Processing (BSP). It has important theoretical and applied value. It has attractive application prospects in widely fields, such as telecommunications, audio signal separation, biomedical signal processing, and image processing. It is playing an increasingly important role.A new feature extraction method----Independent Component Analysis is used in the PET image denoising in the dissertation. The major work is as following: First, summarize history of medical image, basic theory and features of medical image device----PET, and research status of related medical image processing; Secondly, expound basic theory, restrict conditions, history and research status of ICA, and introduce the implementation process of ICA algorithm in details. Thirdly, pretreat the experimental data of PET image by PCA algorithm. Then extract features of the pretreated data by FastICA algorithm based on Negentropy. Denoise the features extracted by Sparse Code Shrinkage algorithm. Restruct the original PET image. Finally, compare and analyze the method in this dissertation and median filter and wave filter during the process of PET image denoising. The experimetal result proved that the denoising algorithm used in the dissertation is applicable to PET image compared to other algorithm commonly used in medical image. And the experimental result is very good.