Font Size: a A A

The Algorithm Research Of Hepatic Portal Vein Segmentation Based On CT Images

Posted on:2016-07-29Degree:MasterType:Thesis
Country:ChinaCandidate:S L CaoFull Text:PDF
GTID:2348330479454691Subject:Computer technology
Abstract/Summary:PDF Full Text Request
With the rapid development of medical imaging technology, the analysis of the liver vascular imaging has become a routine examination method in liver clinical examination, diagnosis, surgical navigation. The quickly and accurately segmentation on the liver vascular images can not only provide the geometry, physics, statistical parameters of the liver vascular structure, but also provide the basis for the diagnosis and treatment process for doctor to specify the timely and accurately treatment plan.Different from the traditional semi-automatic blood vessels segmentation method, we use automatic portal vein trunk locating method to extract the portal vein trunk area, and then use the 3D region-grow method to segment the vessels rather than the traditional method, without human intervention and automate the segmentation process.Automatic hepatic portal vein segmentation method is mainly divided into CT data preprocessing, vascular enhancement, hepatic portal vein segmentation. In CT data preprocessing phase, first, we use the anisotropic diffusion filter to weaken noise and extract the bone area, reducing the influence of image quality and noise in the segmentation results. After the preprocessing, we introcude a multi-scale Hessian vascular enhancement filter to enhance blood vessels in the volume data. Finally, in the hepatic portal vein segmentation phase, we firstly introduce a automatic method to locate the spine with prior knowledge of anatomy to accomplish the portal vein trunk detection on the basis of the spine, then use the 3D region-grow method with the seed point to accomplished the segmentation. Finally, vessel repairing on the basis of the preliminary segmentation result makes the result more accurate.The effectiveness of our method is demonstrated by experimental result on different sequences of abdominal CT images. Experiments show that our proposed algorithm is capable of accurately, automatically, effectively extracted the hepatic portal vein vessel.
Keywords/Search Tags:Vessel Segmentation, Portal Vein Trunk Location, Multi-scale Hessian Vascular Enhancement, Vessel Repairing
PDF Full Text Request
Related items