Localized Region Based Active Contour Algorithm for Segmentation ofAbdominal Organs and Tumors in Computer Tomography Images

摘要

This study presents a localized region based lankton algorithm with decision median filter for thesegmentation of abdominal organs and tumors in CT images. Active contour algorithms based on discreterepresentation of implicit functions are widely used in medical image segmentation. The decision median filtercan filter the gaussian noise and impulse noise efficiently than conventional spatial domain filters with betteredge preservation. A comparative analysis of various active contour models is done and the decision medianfilter based lankton segmentation algorithm gives best result. The performance of the proposed segmentationalgorithm was evaluated in terms of metrics like dice coefficient and hausdorff distance. The proposed algorithmeffectively segments the abdominal organs and tumors in computer tomography images with well-definedboundaries.