改进区域生长法的肝部CT图像ROI提取-《计算机技术与发展》

文章信息/Info

作者:: 李仔麒; 马慧彬; 李殿奎; 范蕊; 佳木斯大学信息电子技术学院,黑龙江佳木斯,154007;佳木斯大学信息电子技术学院,黑龙江佳木斯 154007;佳木斯大学整合医学研究院,黑龙江佳木斯 154007

Author(s):: LI Zi-qi 1 ; MA Hui-bin 1; 2 ; LI Dian-kui 1; 2 ; FAN Rui 1; 1. School of Information Technology,Jiamusi University,Jiamusi 154007,China;2. Integrated Medical Research School,Jiamusi University,Jiamusi 154007,China

Keywords:: liver CT image; ROI; threshold segmentation method; image segmentation; regional growth method; morphology

摘要:: 为了能够将肝部CT图像ROI进行准确的分割,并解决阈值分割法区域不精准、时间复杂度高以及传统区域生长法的噪声干扰和灰度不均而出现的图像过分割和空洞问题,提出了一种结合形态学的改进区域生长算法.改进区域生长算法充分利用了传统区域生长算法良好的分割效果和边界信息保持等优点,并用数学形态学的腐蚀与膨胀等相关技术弥补了传统区域生长算法造成的图像过分割和边缘不平滑问题.有效地解决了阈值分割法在分割图像时出现的分割不准确、用时较长的问题.根据仿真实验中三组图片的结果表明,改进区域生长法不仅能实现对肝部CT图像的ROI精准分割,很好地保留了边缘信息,而且降低了算法的时间复杂度,图像中的空洞问题得到了很好地解决,有助于肝部CT图像的识别和分类.

Abstract:: In order to accurately segment the ROI of liver CT images and solve the problem of over-segmentation and void caused by re-gional inaccuracy,high time complexity,noise interference and gray inconsistency of traditional regional growth methods,we propose animproved regional growth method combining morphology. It makes full use of the advantages of better segmentation effect and boundaryinformation maintenance of traditional region growth method and makes up the image over-segmentation and edge smoothing caused bytraditional region growth method by using mathematical morphological corrosion and expansion techniques. The problem of image seg-mentation inaccuracy and time-consuming is solved effectively. According to the results of three groups of pictures in the simulation,theimproved region growth method can not only realize the ROI accurate segmentation of the liver CT image,preserving the edge informa-tion well,but also can reduce its time complexity. The problem of hole can be solved well in the image,which is helpful for the recogni-tion and classification of the liver CT image.