Automatic segmentation of liver tumor on computed tomography images
Megha P. Arakeri, Lakshmana · 2010
Segmentation of liver tumor on Computed Tomography (CT) images is a challenging task due to anatomic complexity and the imaging system noise. The conventional region growing method has a widespread use in medical image segmentation because of its robustness to noise. However, region growing algorithm is semi-automatic in which the initial seed point and threshold value have to be manually identified. To avoid these problems, in this paper we propose a automatic region growing method that incorporates fuzzy c-means clustering algorithm to find the threshold value and modified region growing algorithm to find seed point automatically. In this paper, we also describe a framework to create a three dimensional (3D) model of the liver which can be used by the surgeons for tumor volume measurement, liver transplant and surgical planning. The proposed method has been tested on several CT images of liver. The results show that the algorithm successfully detects the edges of the liver tumor distinguishing it from the background without manual intervention.