基于Graph Cuts的脑部MRI图像脑组织提取方法

doi:10.3969/j.issn.0258-8021. 2014. 05.02

摘要
图/表
参考文献
相关文章 (0)

全文: PDF (1074 KB) HTML (1 KB)
输出: BibTeX | EndNote (RIS)

摘要脑部MRI图像的脑组织提取是神经影像学分析的一项重要预处理过程，为提高提取精度，提出一种基于graphcuts的脑组织自动提取方法，主要适用于T1加权MRI图像。首先采用Smith等提出的脑组织提取工具（BET）得到感兴趣区域 (ROI)，仅在该区域内用graphcuts方法进行演化；并在graphcuts中加入一个速度限制因子，解决脑组织提取过程中的局部收敛和边界泄漏问题；此外，还采用一种逐层处理2D图像切片的3D数据初始化方法。利用IBSR(Internet Brain Segmentation Repository)网站提供的18组数据，将所提出方法与现有的部分脑组织提取方法（脑组织提取工具(BET)、脑组织表面提取算法(BSE)、分水岭算法(WAT)、混合分水岭算法(HWAT)、图割算法(GCUT) 和鲁棒脑组织提取算法(ROBEX)），进行对比试验。结果显示，本方法最接近于标准分割，平均Dice系数达到095，并且在多个评价参数（假阳性率32%和Hausdorff距离96）上都取得最好结果。实验表明，所提出方法具有较好的精确性和稳定性。

	服务

	把本文推荐给朋友
	加入我的书架
	加入引用管理器
	E-mail Alert
	RSS
	作者相关文章
	杨素华^*陈琼罗艳芬

关键词 ：脑组织提取, graph cuts, BET方法, 感兴趣区域

Abstract：The brain extraction from cerebral T1 weighted MRI volume is an important preprocedure for neuroimage analysis. To improve the precision of extraction, an automatic brain extraction method based on a graphcuts model was proposed. The method uses the brain extraction tool (BET) to obtain the region of interest (ROI) and only performs graphcuts model in the ROI. A velocity factor was added in the graphcuts model to eliminate the boundary leakage and local convergence. A slice by slice contour initial method was also used to initialize the 3D brain boundary. The method was applied to 18 cerebral MRI volumes provided by the Internet Brain Segmentation Repository (IBSR). In testing, the mean Dice similarity coefficient was 095 and the results obtained by our method were very similar to those produced by manual segmentation and achieved the best results on many of the evaluation metrics (false positives rate 32% and Hausdorff distance 96) for the IBSR data in comparison of our method with existing brain extraction methods including brain extraction tool (BET), brain surface extraction (BSE), watershed algorithm (WAT), hybrid watershed algorithm (HWAT), graphcuts (GCUT) and robust brain extraction (ROBEX). The experiment showed the proposed method was precise and robust.

Key words： brain extraction graph cuts BET region of interest

基金资助:国家自然科学基金（61162023，61163046）

引用本文:

杨素华^*陈琼罗艳芬. 基于Graph Cuts的脑部MRI图像脑组织提取方法[J]. 中国生物医学工程学报, 2014, 33(5): 525-531.
YANG Su Hua^*CHEN Qiong LUO Yan Fen. Automatic Brain Extraction Method from Cerebral MRI Image Based on GraphCuts. journal1, 2014, 33(5): 525-531.

链接本文:

http://cjbme.csbme.org/CN/10.3969/j.issn.0258-8021. 2014. 05.02 或 http://cjbme.csbme.org/CN/Y2014/V33/I5/525

［1］Smith SM. Fast robust automated brain extraction ［J］. Hum Brain Mapp, 2002, 17(3): 143-155.
［2］Shattuck DW, Sandorleahy SR, Schaper KA, et al. Magnetic resonance image tissue classification using a partial volume model ［J］. Neuroimage, 2001, 13(5): 856-876.［3］Hahn HK, Peitgen HO. The skull stripping problem in MRI solved by a single 3D watershed transform ［J］. Lect Notes Comput Sci, 2000, 1935: 134-143.
［4］Ségonne F, Dale AM, Busa E, et al. A hybrid approach to the skull stripping problem in MRI ［J］. Neuroimage, 2004, 22(3): 1060-1075.
［5］Huang A, Abugharbieh R, Ram R, et al. MRI brain extraction with combined expectation maximization and geodesic active contours ［C］ // Rabab W, Fayez G, eds. The 6th IEEE International Symposium on Signal Processing and Information Technology. Vancouver: IEEE, 2007: 107-111. 
［6］Liu Jiaxiu, Chen Yongsheng, Chen Lifen. A accurate and robust extraction of brain regions using a deformable model based on radial basis functions ［J］. Journal of Neuroscience Methods, 2009, 183(2): 255-266.
［7］Wang Yaping, Nie Jingxin, Yap PT, et al. Robust deformable surface based skull stripping for largescale studies ［C］ // Peters T, Fichtinger G, Martel A, eds. Medical Image Computing and ComputerAssisted Intervention (MICCAI 2011). Toronto: Springer Berlin/Heidelberg, 2011: 635-642.
［8］Leung KK, Barnes J, Modat M, et al. Brain MAPS: an automated, accurate and robust brain extraction technique using a template library ［J］. Neuroimage, 2011, 55(3): 1091-1108.
［9］Eskildsen SF, Coupé P, Fonov V, et al. BEaST: brain extraction based on nonlocal segmentation technique ［J］. Neuroimage, 2012, 59(3): 2362-2373.
［10］税午阳，周明全，耿国华. 磁共振颅脑图像的脑组织自动获取方法［J］. 软件学报, 2009, 20(5): 1139-1145. 
［11］贾迪，杨金柱，张一飞，等. 序列磁共振颅脑影像的脑组织自动提取方法［J］. 仪器仪表学报, 2011, 32(8): 1781-1787.
［12］韩翀蛟，林相波，马慧超，等. 基于层间先验知识从脑MRI图像中自动提取脑组织［J］. 生物医学工程与临床, 2011, 15(2): 111-115
［13］江少锋，王文辉，陈武凡，等. 基于改进BET算法的MR颅脑图像脑组织自动提取［J］. 中国图象图形学报, 2009, 14(10): 2029-2034.
［14］江少锋，万红平，陈震，等. 利用感兴趣区域从脑部MRI中提取脑组织［J］. 中国图象图形学报, 2013, 18(12): 1644-2034.
［15］Boykov Y, Jolly MP. Interactive graph cuts for optimal boundary & region segmentation of objects in ND images ［C］ // International Conference on Computer Vision (ICCV 2001). Vancouver: IEEE, 2001，1: 105-112.
［16］Sadananthan S, Zheng Weili, Chee M, et al. Skull stripping using graph cuts ［J］. Neuroimage, 2010, 49(1): 225-239.
［17］Iglesias JE, Liu CY, Thompson PM, et al. Robust brain extraction across datasets and comparison with publicly available methods ［J］. IEEE Trans Med Imaging, 2011, 30(9): 1617-1634.
［18］Harvard Medical School. IBSR数据［DB/OL］. https://www.nitrc.org/projects/ibsr, 2013-07-05/2013-10-15.