Research On Tumor Segmentation Of Liver CT Image Based On Region Growing

Abdominal CT image is the most common imaging technique for liver tumor examination.As a computer-aided technology,it can help doctors get the tumor size,number,location and other information.The segmentation of liver tumor based on abdominal CT images is to mark out the liver tumor from the image and analyze the segmentation results,to provide a basis for the doctors who decide the follow-up clinical treatment.Therefore,it will be of great importance to segment the image of liver tumor from the abdominal CT.This paper first studies the characteristics of abdominal CT images and analyzes the commonly used medical image segmentation algorithms.The regional growth algorithm is selected as the image segmentation method.Because the image segmentation using region growing algorithm can make the segmentation region in the space is tightness.Therefore,this method is suitable for the segmentation of structural images,such as tumors or wounds with uniform internal features.Then,the paper analyzes the principle,advantages and disadvantages of traditional regional growth algorithm,and finds that it has two problems in CT image segmentation of liver tumor:The one is to choose the initial seed point position.by artificial manual selection of seed points,it is time-consuming and labor-intensive,and the method is not easy to popularize the use.The other is to make the ruler of growing.The quality of the growing ruler directly influences the segmentation results.Aiming at the above problems,this paper proposes an improved region growing algorithm for automatically to select seed points,automatically to segment liver and tumor from abdominal CT images.Firstly,the improved algorithm,integrating the image preprocessing method,select the inscribed circle center of the largest connected region in the abdominal CT image as the seed point.Secondly,the growth rule is improved and optimized to reduce the void in the segmented image,and to realize the accurate segmentation of the liver by the improved algorithm.Finally,with morphological method,the segmentation liver is to reprocessed,to extract the approximate tumor area,region growing algorithm is secondly used to the tumor is segmented from the image.The improved algorithm is proved to be effective and accurate by the relevant experiments.