Research On The Methods Of Neck Tissue Extracted From CTA Images

In recent years, the medical imaging technology has been one of the important technical means for the modern medical diagnosis. With the computer-aided diagnostic systems and medical images segmentation methods, the interesting regions can be extracted accurately and quickly, which can improve the reliability of clinical diagnosis. Computerized Tomography Angiography (CTA) has been widely used in medical image analysis. Comparing with other similar technologies, the bloods and bones in the CTA images have highlighted. Consequently, CTA images have higher clinical diagnosis in checking vascular and bone diseases.The neck tissues of CTA images suffer from asymmetrical grayscales and noise, the borders and shapes are fuzzy. Thus, the segmentation of neck tissues is a difficult task. Allowing for the demand of clinical diagnosis and many diseases rooted in the vessels and bones, the thesis studies the algorithms of vessels and vertebrae extracted from the neck CTA images. The main part of the thesis includes the following:First of all, on the basis of the medical image segmentation methods and analyzing the specific features of the head and neck in CTA images, this thesis designs the pre-segmentation which allows for the task needs. This part can establish the foundation for subsequent segmentation and mainly includes:removing noise based on the improved PM diffusion model, extracting the head and neck images and roughly extracting the vertebra areas. Secondly, this thesis designed and implemented the reconstruction algorithm for the three-dimensional vessels based on the region circularity and grayscales. This method can extract complete and smooth vessels, and doesn’t need the adherent segmentation compared with the traditional methods. It can improve the efficiency and the quality of the vessels’ extraction. Finally, this thesis designed the vertebra extraction algorithm based on the improved C-V model and scan line filling method. It can solve the missed of the spinal cord and vertebra can’t be extracted completely. All of the algorithms were designed and implemented in the Neusoft MIC application. The result of the experiment indicated that all the algorithms can obtain the vessels and vertebra accurately and quickly. The algorithms have certain values for the application of computer-aided diagnostic system.