期刊名称：International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
印刷版ISSN：2278-1323
出版年度：2017
卷号：6
期号：1
页码：114-118
出版社：Shri Pannalal Research Institute of Technolgy
摘要：Abdomen/Liver cancer is the rapid creation ofabnormal cells that grow beyond their usual boundaries andspread to other organs. Computer aided technique helps toidentify the shape and the distribution of cells of tumorautomatically without human intervention resulting in earlydiagnosis. The work in this paper presents an image processingpipeline, consisting of noise reduction, feature extraction byimage segmentation, morphological operations to extract theregion of interest i.e. tumor. Two segmentation techniques,namely, region growing and Otsu segmentation areimplemented to extract the area of abdomen/liver containingtumor from the CT images. The performance of both themethods performed on local data set of CT liver images iscompared with the ground truth and it has been observed thatthe region growing segmentation technique yielded betterresults. An accuracy of 87.5% with specificity and sensitivity of66.66% and 94.44 respectively has been observed for regiongrowing method.
