Objective To segment the CT images of liver neoplasms accurately. Methods The image was preprocessed through MATLAB firstly, then was successfully segmented with the methods of gray transferring, binaryzation processing, inverting, morphological treatment and the region-growing method. Results The experimental result showed that the combination segmentation was simple, rapid, and particularly suitable for small lesions, and liver neoplasms could be accurately segmented. Conclusions The combination segmentation was effective to some extent in liver neoplasms, however was limited in the tumors with massive adhesion of surrounding tissues.
|