Subcategory:
Category:
Words:
344Pages:
1Views:
184Liver is most commonly involved by metastatic diseases due to its enriched connectivity with the blood vessels and lymphatic vessels being liver cancer one of the leading causes of death worldwide Tumor volume measurement accurate staging analysis tumors features such as site local spread distant and involvement with other organs in the body and early detection are decisive to improve the curability and cancer follow up helping the physician to decide what type of treatment is best suited and to predict patient s prognosis In the field of medical diagnosis an extensive diversity of imaging techniques have been incorporated such as ultrasound imaging computed tomography CT and magnetic resonance imaging MRI The development of image processing and segmentation techniques plays an important role in allowing the extraction of metastatic liver tumor from the surrounding tissue Manual segmentation is time consuming prone to errors dependent on users and requires expert knowledge to yield accurate and robust results Therefore CAD Computer Aided Diagnosis oriented diagnosis has become one of the most extensive research fields
The main target in this is to obtain the enhanced tumor boundary This is because the edges of the tumor are not clear and blurs away and our segmentation method performs worse on such conditions The EM MPM algorithm gives the most efficient Hepatic CT pre processing results when the tumor is located inside the periphery and in case of multiple tumors at a distant location from each other which can be used to further segment the tumor After the processing the segmentation part was implemented It involved extraction of the tumor boundary by using traditional snake model The main target in this was to initialize the points near the tumor boundary and converge the snake to the boundary of the tumor using energy constraints By keeping the number of iterations 150 α 0 1 β 0 05 the metastatic regions with no major issues with concavity are successfully delineated or else region based segmentation is implemented in the case of concavities where the tumor is located inside the periphery
The main difficulties in tumor segmentation include ambiguity between liver tissue and tumors boundaries complexity of tumors surfaces contrast variability between liver parenchyma tumors and vessels lesions can be brighter hyper dense or darker hypo dense than the surrounding tissue different tumor size shape and location patient conditions presence of neighbouring structures organs with the same density values and possible presence of many small metastases located very close to each other Due to these multiple difficulties medical and image processing researches are going on extensively in order to extract the metastatic tumor boundaries from the Hepatic CT images efficiently The implementation of the EM MPM algorithm and segmentation techniques during the project shows good results on segmenting the tumors inside the periphery and at distant location apart when applied to segment the metastatic tumors from the 15 Dataset of Hepatic CT which includes both single tumor and multiple tumors containing sets The project successfully implements the semi automatic way of CAD Computer Aided Diagnosis oriented pre processing and segmentation of the metastatic tumor