Abstract

• The objective of this research is the liver lesion regions detection and extraction for diagnostics. It involves four stages in this work such as pre-processing, extraction of liver, detection of liver, segmentation of liver and estimation of lesion regions. The outputs of lesion regions are analyzed from segmented livers. Lesion regions detection stage is used with mean-multi thresholding method and implemented with Matlab Programming language. Hundred data sets of CT abnormal liver images (256  256 pixel, 2D liver slices) are tested and lesion regions are calculated from the criteria that the number of gray pixels which are greater than mean value of that block is greater than one-third time of total pixels. The comparative results in this research can be shown with mean value of lesion regions of benign liver CT images is in the range 149-167 and malignant liver CT images is in the range 125-135.