期刊名称：International Journal of Image Processing (IJIP)
电子版ISSN：1985-2304
出版年度：2011
卷号：5
期号：2
页码：166-176
出版社：Computer Science Journals
摘要：Liver lesion segmentation using single threshold in 2D abdominal CT images proves insufficient. The variations in gray level between liver and liver lesion, presence of similar gray levels in adjoining liver regions and type of lesion may vary from person to person. Thus, with threshold based segmentation, choice of appropriate thresholds for each case becomes a crucial task. An automatic threshold based liver lesion segmentation method for 2D abdominal CT pre contrast and post contrast image is proposed in this paper. The two thresholds, Lower Threshold and Higher Threshold are determined from statistical moments and texture measures. In pre contrast images, gray level difference in liver and liver lesion is very feeble as compared to post contrast images, which makes segmentation of lesion difficult. Proposed method is able to determine the accurate lesion boundaries in pre-contrast images also. It is able to segment lesions of various types and sizes in both pre contrast and post contrast images and also improves radiological analysis and diagnosis. Algorithm is tested on various cases and four peculiar cases are discussed in detail to evaluate the performance of algorithm.
