基于水平集的医学图像分割算法

doi:10.16180/j.cnki.issn1007-7820.2021.02.003

摘要/Abstract

摘要：

针对具有灰度不均效应的医学图像分割问题,文中提出了一种快速结合全局和局部信息的水平集模型,即HLSGL。基于C-V水平集分割算法的全局信息,通过计算图像各像素点的局部拟合均值,引入局部信息能量项,使全局与局部信息叠加构成驱动力项,保证对图像边缘具有较好的局域化效果。在构造的驱动力项中引入一种新的速度停止函数,使分割过程能自适应地调节曲线演化速率,提高了分割效率。将HLSGL模型应用于不同种类医学图像,实验表明该方法可以高效地分割含噪声、弱边界、灰度不均的医学图像,得到较完整的轮廓曲线。与其他水平集模型的对比实验表明,HLSGL模型的准确性、鲁棒性、分割效率均得到改善。

关键词: 医学图像分割, 灰度不均, 水平集, C-V模型, 全局和局部, 速度停止函数

Abstract:

In view of the common inhomogeneous intensity effect in medical images, a hybrid level set model combining global and local term is proposed, namely HLSGL. Based on the global information of the C-V level set segmentation algorithm, the local information energy term is introduced by calculating the local fitting average of each pixel of the image, so that the global and local information can be superimposed to form the driving force term, ensuring better localization effect on the edge of the image. A new speed stopping function is introduced into the driving force term to adaptively adjust the curve evolution rate in segmentation process, improve segmentation efficiency. In addition, HLSGL is applied to different kinds of medical images in the present study, and experiments results show that the HLSGL model can efficiently segment medical images with noise, weak boundaries and inhomogeneous intensity to obtain a more complete contour curve. Compared with other level set models,the accuracy, robustness and segmentation efficiency of the HLSGL model are significantly improved.

Key words: medical image segmentation, inhomogeneous intensity, level set, C-V model, global and local, speed stopping function

中图分类号:

TP391.41

房巾莉,吕毅斌,王樱子,唐胜男,武德安. 基于水平集的医学图像分割算法[J]. 电子科技, 2021, 34(2): 12-20.

FANG Jinli,LÜ Yibin,WANG Yingzi,TANG Shengnan,WU Dean. A Novel Medical Image Segmentation Algorithm Based on Level Set[J]. Electronic Science and Technology, 2021, 34(2): 12-20.

图/表 17

图1

图2

图3

表1

图3的迭代次数和分割时间的对比"

	葫芦图		细胞图
	Iteration	Time	Iteration	Time
无h $x$	10	29.42	90	25.66
有h $x$	7	26.43	60	22.60

表1

图4

表2

图4的分割效率对比"

	合成图		大米图
	Iteration	Time	Iteration	Time
选取g $x$	400	99.74	100	51.78
选取h $x$	130	27.20	8	20.17

表2

图5

图6

表3

图7

表4

图8

表5

图9

图10

图11

表6

参考文献 25

[1]	何菁, 陈胜. 一种全新的两步自动化医学图像分割方案[J]. 电子科技, 2016,29(7):85-87.
	He Jing, Chen Sheng. A new two-step automatic segmentation scheme for medical images[J]. Electronic Science and Technology, 2016,29(7):85-87.
[2]	兰红, 韩纪东, 方毅. 活动轮廓模型在医学图像分割的综述[J]. 科学技术与工程, 2018,18(16):164-170.
	Lan Hong, Han Jidong, Fang Yi. A survey of active contour model for medical image segmentation[J]. Science Technology and Engineering, 2018,18(16):164-170.
[3]	贝琛圆, 于海滨, 潘勉, 等. 基于改进U-Net网络的腺体细胞图像分割算法[J]. 电子科技, 2019,32(11):18-22.
	Bei Chenyuan, Yu Haibin, Pan Mian, et al. Gland cell image segmentation algorithm based on improved U-Net network[J]. Electronic Science and Technology, 2019,32(11):18-22.
[4]	Liu Z W, Li X X, Luo P, et al. Deep learning markov random field for semantic segmentation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2018,40(8):1814-1828. doi: 10.1109/TPAMI.2017.2737535 pmid: 28796610
[5]	Le T H N, Quach K G, Luu K, et al.Reformulating level sets as deep recurrent neural network approach to semantic segmentation[J].IEEE Transactions on Image Processing, 2018(9):11-18.
[6]	Kiechle M, Storath M, Weinmann A, et al. Model based learning of local image features for unsupervised texture segmentation[J]. IEEE Transactions on Image Processing, 2018,27(4):1994-2007.
[7]	Caselles V, Catt é F, Coll T, et al. A geometric model for active contours in image processing[J]. Numerical mathematic, 1993,66(1):1-31.
[8]	Malladi R, Sethian J A, Vemuri B C. Shape modeling with front propagation:a level set approach[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1995,17(2):158-175.
[9]	Osher S, Sethian J A. Fronts propagating with curvature-dependent speed: algorithms based on Hamilton-Jacobi formulations[J]. Journal of computational physics, 1988,79(1):12-49.
[10]	Tony F Chan, Luminita A Vese. Active contours without edges[J]. IEEE Transactions on Image Processing A Publication of the IEEE Signal Processing Society, 2001,10(2):266-277. pmid: 18249617
[11]	Li C, Kao C Y, Gore J C, et al. Implicit active contours driven by local binary fitting energy [C].Minneapolis:IEEE Conference on Computer Vision and Pattern Recognition, 2007.
[12]	Li C M, Kao C Y, Gore J C, et al. Minimization of region-scalable fitting energy for image segmentation[J]. IEEE Transactions on Image Processing, 2008,17(10):1940-1949. pmid: 18784040
[13]	Wang X F, Huang D S, Xu H. An efficient local Chan-Vese model for image segmentation[J]. Pattern Recognition, 2010,43(3):603-618.
[14]	Caselles V, Kimmel R, Sapiro G. Geodesic active contours[J]. International Journal of Computer Vision, 1997,22(1):61-79.
[15]	Lankton S, Tannenbaum A. Localizing region-based active contours[J]. IEEE Transactions on Image Process, 2008,17(11):2029-2039.
[16]	Shi Y, Karl W C. A real-time algorithm for the approximation of level-set based curve evolution[J]. IEEE Transactions on Image Process, 2008,17(3):645-656.
[17]	Li C M, Xu C Y, Gui C F, et al. Distance regularized level set evolution and its application to image segmentation[J]. IEEE Transactions on Image Processing, 2010,19(12):3243-3254. doi: 10.1109/TIP.2010.2069690 pmid: 20801742
[18]	Zhang W H, Wang X, You W, et al. RESLS: Region and edge synergetic level set framework for image segmentation[J]. IEEE Transactions on Image Processing, 2019,29(1):57-71.
[19]	Li C M, Xu C Y, Gui C F. Level set evolution without re-initialization: a new variational formulation[J]. Computers and Signal Processing, 2005,51(1):25-27.
[20]	Vovk U, Pernus F, Likar B. A review of methods for correction of intensity inhomogeneity in MRI[J]. IEEE Transactions on Medical Imaging, 2007,26(3):405-421. doi: 10.1109/TMI.2006.891486 pmid: 17354645
[21]	Nair J J, Govindan V K. Intensity inhomogeneity correction using modified homomorphic unsharp masking[J]. Journal of Medical Imaging and Health Informatics, 2014,4(2):285-290.
[22]	闻敬东, 张轩雄, 巨志勇. 非均匀图像亮度补偿的分割方法[J]. 电子科技, 2017,30(9):12-15.
	Wen Jingdong, Zhang Xuanxiong, Ju Zhiyong. Segmentation method for unevenly image brightness compensation[J]. Electronic Science and Technology, 2017,30(9):12-15.
[23]	Yang Y Y, Jia W J. Efficient and robust segmentation and correction model for medical images[J]. IET Image Processing, 2019,13(12):2245-2254.
[24]	Dietenbeck T, Alessandrini M, Friboulet D, et al. Creaseg: A free software for the evaluation of image segmentation algorithms based on level-set [C].Hongkong:Proceedings of the International Conference on Image Processing, 2010.
[25]	Khadidos A, Sanchez V, Li C T. Weighted level set evolution based on local edge features for medical image segmentation[J]. IEEE Transactions on Image Processing, 2017,26(9):1979-1991.

	血管图1	血管图2	血管图3	手指图
Size	110×111	131×103	183×189	148×110
λ₁	1	1	0.1	1
λ₂	10	10	18	10
μ	0.01	0.01	0.01	0.01
Numiter	50	45	160	147
CPU time /s	41.51	38.00	60.87	53.78

	肾囊图	肝囊图	脑肿瘤图
Size	114×154	101×140	174×153
λ₁	1	0.1	1
λ₂	1	1	1
μ	0.01	0.01	0.01
Numiter	25	15	11
CPU time/s	15.27	14.20	14.16

		CV	LBF	RSF	LCV	HLSGL
子	DSC	0.864	0.893	0.335	0.879	0.952
宫	Numiter	100	500	30	160	50
囊	Time	31.06	167.22	12.05	55.92	25.44
腕	DSC	0.875	0.546	0.834	0.871	0.987
关	Numiter	50	500	100	40	40
节	Time	29.620	166.910	16.079	23.390	28.180
骨	DSC	0.861	0.699	0.423	0.952	0.981
裂	Numiter	15	90	100	110	300
图	Time	16.360	36.370	23.930	40.101	90.840

		Caselles	CV	Lankton	Shi	HLSGL
肺部	DSC	0.326	0.903	0.431	0.956	0.983
	Numiter	400	50	300	300	50
	Time	344.92	113.69	91.57	461.33	89.96
血管	DSC	0.369	0.853	0.316	0.897	0.968
	Numiter	300	150	50	500	40
	Time	16.35	36.97	1.26	308.09	31.45