计算机工程与应用 ›› 2022, Vol. 58 ›› Issue (14): 200-209.DOI: 10.3778/j.issn.1002-8331.2011-0281

• 图形图像处理 • 上一篇    下一篇

改进DRLSE的分步式肝脏及肿瘤分割方法

郭贝贝,马自萍,王兴岩,杨珂   

  1. 1.北方民族大学 数学与信息科学学院,银川 750021
    2.北方民族大学 计算机科学与工程学院,银川 750021
  • 出版日期:2022-07-15 发布日期:2022-07-15

DRLSE-Based Split-Step Liver and Tumor Segmentation Algorithm

GUO Beibei, MA Ziping, WANG Xingyan, YANG Ke   

  1. 1.School of Mathematics and Information Science, North Minzu University, Yinchuan 750021, China
    2.School of Computer Science and Engineering, North Minzu University, Yinchuan 750021, China
  • Online:2022-07-15 Published:2022-07-15

摘要: 为了提升CT图像肝脏及肝脏肿瘤的分割精度,提出一种改进DRLSE的分步式肝脏及肿瘤分割方法。第一阶段:采用分步式分割方法对肝脏进行分割,(1)采用阈值处理、形态学方法、自适应区域生长方法进行肝脏的粗分割;(2)采用数学形态对分割结果进行优化,进行肝脏的细分割。第二阶段:构造参数梯度形态学和各向异性扩散滤波的距离正则化水平集演化(改进的DRLSE)模型进行肝脏肿瘤分割。实验利用3Dircadb数据集验证方法的有效性,计算了DICE、VOE、ASD和MSD指标评估分割的性能。实验结果表明该方法无需进行训练过程和统计模型的建立,对于复杂的形状和强度变化的CT图像分割效果尤为明显。由定量分析的数值结果显示,分割性能均优于比较算法,提高了分割准确率,具有较强的鲁棒性,为医生诊断和治疗肝癌提供帮助。

关键词: 水平集, 区域生长算法, 距离正则化水平集演化(DRLSE), 三维重建, 各向异性扩散滤波

Abstract: In order to improve the segmentation accuracy of liver and liver tumors in CT images, a stepwise segmentation method with improved DRLSE is proposed. The first stage: the liver is segmented by the stepwise segmentation method.  (1) The liver is coarse-segmented by threshold processing, morphological method and adaptive region growth method.  (2) Mathematical morphology is used to optimize the segmentation results for liver segmentation. The second stage:it constructs the distance regularized level set evolution (improved DRLSE) model of parametric gradient morphology and anisotropic diffusion filtering for liver tumor segmentation. In the experiment, the validity of the method is verified by using 3Dircadb data set, and the performance of DICE, VOE, ASD and MSD indexes are calculated to evaluate the segmentation accuracy.  The experimental results show that this method does not require the establishment of training process and statistical model, and is particularly effective for CT image segmentation with complex shape and strength changes. The numerical results of quantitative analysis show that the segmentation performance is superior to the comparison algorithm, improves the segmentation accuracy, and has strong robustness, which provides help for doctors to diagnose and treat liver cancer.

Key words: level set, region growing, distance regularized level set evolution(DRLSE), 3D reconstruction, anisotropic diffusion filtering