Pixel Level Fusion Research Based On Non-small Cell Lung Cancer PET/CT Image

Background PET/CT fusion image has important application value for image analysis and clinical diagnosis. But the current studies in PET/CT image fusion are mostly based on clinical applications, but it has few studies in algorithms which are most based on wavelet transform.Objectives Non-small cell lung cancer PET images and CT images regard as research objects. Fusing PET and CT images by combination dual-tree complex wavelet transform and membership function to get richer medical information of non-small cell lung cancer PET/CT fusion images. It provides better basis for medical diagnosis and improves the accuracy of diagnosis.Methods Two fusion algorithms are proposed by combining dual-tree complex wavelet transform and membership function. The first is a fusion algorithm of PET/CT based on dual-tree complex wavelet transform and self-adaption Gaussian membership function, and the second is a self-adaption fusion algorithm of PET/CT based on DTCWT and the combination of membership functions. The two algorithms are using for fusing non-small cell lung cancer PET images and CT images. Subjective and objective methods are using for evaluating the effect of fusion images.Results For the first fusion algorithm three experiments were done, comparison experiment of the other pixel-level fusion algorithms, the fusion effect objective evaluation experiment and comparative experiment with different fusion rules of dual-tree complex wavelet transform. The experiment results show that the algorithm can improve the information entropy by 7.23%, and mutual information by17.98%. That is to say the algorithm is an efficient fusion method.For the second fusion algorithm two experiments are done, comparison experiment of the other pixel-level fusion algorithms and the experiment of fusion effect objective evaluation. The experimental results show that the algorithm can better retain and show the edge and texture information of lesions. Then, in order to verify the validity and feasibility of the fusion algorithm objectively, we used 32 pieces of non-small cell lung cancer PET images and CT images to do the simulates experiment. The final results show that the algorithm is superior to other pixel level fusion algorithms from both the subjective and objective evaluation.