Multimodal Miniature Pig Heart Image Segmentation And Registration

Ischemic heart disease has become the first cause of death disease in the world, Medical imaging technology is an important technical means for accurate diagnosis of ischemic heart disease, and optimize treatment regimens and efficacy assessments. Basic research of new treatments is necessary utilization of several cross-modality medical imaging technology, comprehensive and accurate assess cardiovascular structure and cardiac function from the function and structure to achieve multiple levels.In this paper, construct miniature pig in vivo ischemic infarction model, using CTA imaging techniques for high-resolution structural information of cardiac imaging; using 18F-FDG imaging performed PET/CT monitor myocardial metabolic function; using RGD targeting angiogenesis PET/CT imaging observe the situation angiogenesis after myocardial infarction. However, due to differences between the image forming apparatus, each modality imaging time is different and animal movement between the imaging device and other reasons caused by different attitude, different modes of image information can not reach the match in space, for a comprehensive observed structural and functional information, we need inter-modal miniature pig heart image segmentation and registration. In this paper, multi-modal imaging evaluation in vivo myocardial infarction model miniature swine, studied the miniature pig model of myocardial infarction CTA image registration and fusion PET/CT problem.Firstly, we complete the single-mode mini pig heart atlas making and provide atlas making process(including CTA map, PET/CT imaging with CT machine map). 18 set of miniature pig model raw data(each data CTA data included six groups, six groups FDG data, six groups RGD data) have been preprocessed in this work. Base on knowledge of cardiac anatomy of the heart has been artificially splited. Using automatic segmentation method based on registration of multiple atlas and each modality atlas library implement miniature pig model of myocardial infarction and PET/CT CTA image in the CT image with modal automatic segmentation.Secondly, base on the results of cardiac automatic segmentation, the paper design image between CTA and PET/CT image registration method, use CTA cardiac atlas and PET/CT cardiac atlas for registration.Use image maps effectively avoid different modal distribution of image gray values are quite different problems, we complete the miniature pig model of myocardial infarction cross-modal registration work.Finally, using cross-modal deformation field information generated by registration, PET/CT function image will be distortion, and then fusion with the CTA image. Completed the integration of structural information and function information. And the final fusion by volume rendering three-dimensional reconstruction of the results, provide technical support for the visual observation of MI state and angiogenesis.In this paper, we achieve small pig heart single-mode image segmentation and cross-modal image registration. Design the heart segmentation and registration software platform for automated processing of image data. Cross-modal fusion results of heart images provides physicians with effective support of medical imaging information for assessment of myocardial ischemia and angiogenesis case.