PET/CT Image Fusion Algorithm Of Lung Cancer Based On Compressed Sensing

Background PET/CT fusion image has important application value for image analysis and clinical diagnosis of lung cancer. At present, the research of PET/CT image fusion is more focused on clinical application, it’s less in algorithm and multi-scale transform is the main method. Medical image fusion based on compressed sensing as the beginning of compressive sensing also is emerging in recent years, it provides technical support for mobile health and wisdom medical.Objectives Lung cancer PET images and CT images regard as research objects. Combining multi-scale transform and compressed sensing used in PET/CT fusion to get richer medical information of 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 multi-scale transform and compressed sensing. One is a fusion algorithm of lung cancer PET/CT based on NSCT and compressed sensing, another is a self-adaption fusion algorithm of lung cancer PET/CT based on compressed sensing and histogram distance. Different simulation experiments are designed to verify these two algorithms, which mainly focuses on subjective evaluation and objective evaluation of fusion results.Results Three comparative experiments were performed on the first fusion algorithm, comparative experiment with different sampling rates, comparative experiment with other image fusion algorithms based on compressed sensing, and with other image fusion algorithms based on multi-scale transform. The experiment results show that compared with other image fusion algorithms based on compressed sensing and multi-scale transform, the standard deviation, average gradient, spatial frequency and mutual information of the algorithm are highest, indicating that the superiority of the algorithm.For the second fusion algorithm three experiments are done, comparative experiment with other image fusion algorithms, comparison of different activity measures and different match measures. The experimental results show that the algorithm could better retain and show the lesion information. Finally, 20 groups of lung cancer PET images and CT images were simulated and the results show that the algorithm is superior to other fusion algorithms in both subjective evaluation and objective evaluation.