[1]徐思敏,金正猛,闵莉花,等.基于测地轮廓和特征函数的灰度异质图像分割[J].计算机技术与发展,2023,33(06):160-167.[doi:10. 3969 / j. issn. 1673-629X. 2023. 06. 024]
　XU Si-min,JIN Zheng-meng,MIN Li-hua,et al.Image Segmentation with Intensity Inhomogeneity Based on Geodesic Contour and Characteristic Function[J].,2023,33(06):160-167.[doi:10. 3969 / j. issn. 1673-629X. 2023. 06. 024]

点击复制

基于测地轮廓和特征函数的灰度异质图像分割()

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:

33

期数:

2023年06期

页码:

160-167

栏目:

人工智能

出版日期:

2023-06-10

文章信息/Info

Title:

Image Segmentation with Intensity Inhomogeneity Based on Geodesic Contour and Characteristic Function

文章编号:

1673-629X(2023)06-0160-08

作者:

徐思敏; 金正猛; 闵莉花; 王 皓; 郭小亚

南京邮电大学 理学院,江苏 南京 210023

Author(s):

XU Si-min; JIN Zheng-meng; MIN Li-hua; WANG Hao; GUO Xiao-ya

School of Science,Nanjing University of Posts and Telecommunications,Nanjing 210023,China

关键词:

图像分割; 灰度异质; 测地轮廓; 交替极小化; 迭代卷积阈值

Keywords:

image segmentation; intensity inhomogeneity; geodesic contour; alternating minimization; iterative convolution threshold

分类号:

TP391

DOI:

10. 3969 / j. issn. 1673-629X. 2023. 06. 024

摘要:

灰度异质图像的分割是图像处理中一项非常有挑战性的任务。 CVB 模型虽然能较好分割灰度异质图像,但是其分割结果容易出现过度分割或欠分割问题。 为了精确分割灰度异质图像,该文在 CVB 模型的基础上,引入基于测地轮廓的长度项来捕捉目标物体的边缘信息,提出一种新的变分分割模型。 同时,为了提高计算效率,该文利用特征函数来表示测地轮廓长度,并且通过基于特征函

数的热核卷积形式逼近测地轮廓的周长。 进一步,结合交替极小化和迭代卷积阈值法,该文设计出一种快速数值求解算法,并且给出了该算法的收敛性和稳定性分析。 最后,对合成图像、核磁共

振图像以及魏茨曼分割数据集上的原始自然图像等三类灰度不均匀图像进行分割实验,并且采用 Dice 相似系数和 Hausdorff 距离作为图像分割的评价指标,实验结果表明:该方法不仅提高了图像分割精度,而且明显提升了收敛速度。

Abstract:

Image segmentation with intensity inhomogeneity is a quite challenging task in image processing. CVB model can be used tosegment images with intensity inhomogeneity well,but?

its segmentation results are prone to over-segmentation or under-segmentation. Inorder to accurately segment images with intensity inhomogeneity, based on the CVB model, a new variational segmentation model isproposed to capture the edge information of the object by introducing geodesic contour length term. At the same time,for improving thecomputational efficiency,we use a characteristic function to represent the geodesic contour length,where the perimeter of geodesic contouris approximated by a heat kernel convolution with the characteristic function. Furthermore,combining the alternating minimization and iterative convolution thresholding method,we design a fast numerical solution algorithm,and the convergence and stability of the algorithmare proved. Finally,segmentation experiments are performed on three kinds of images with intensity inhomogeneity,including syntheticimages,magnetic resonance ( MR) images and original natural images of Weizmann segmentation dataset. Dice similarity coefficient andHausdorff distance are used as evaluation indicators of image segmentation. The experimental results show that the proposed method notonly improves the segmentation accuracy,but also accelerates the convergence speed significantly.

常用功能