﻿ CTA影像头部骨骼组织提取算法
 文章快速检索 高级检索
CTA影像头部骨骼组织提取算法

1. 东北大学 信息科学与工程学院, 沈阳 110819;
2. 东北大学 医学影像计算教育部重点实验室, 沈阳 110819;
3. 沈阳铁路局 信息技术所, 沈阳 110001

Head bone tissue extraction algorithm based on CTA image
CAO Chunhong1,2 , AI Liang3, XU Guangxing1,2
1. College of Information Science and Engineering, Northeastern University, Shenyang 110819, China;
2. Key Laboratory of Medical Image Computing of Ministry of Education, Northeastern University, Shenyang 110819, China;
3. Information Computing Institute, Shenyang Railway Bureau, Shenyang 110001, China
Abstract:Vascular tissue and bone tissue cannot be clearly separated based solely on grayscale information in images of computed tomography angiography (CTA). The algorithm based on the growth of three-dimensional (3D) region improved bone tissue outside the bone contour extraction and the extraction algorithm based on improved Snake model combined with the characteristics of CTA grayscale images were proposed. Combining the knowledge of probability theory to improve the accuracy of determining condition of the region growing, fast skeletal regional seed extraction method of 3D region growing was proposed. It made it possible to obtain more accurate bone tissue area. After the Snake model was selected and some improvements were made to the model, energy image information items were increased, so that the model can better solve the current problems. Finally, the experimental results were given and compared with results from the traditional algorithm. It is confirmed the proposed segmentation of bone tissue extraction algorithm works well.
Key words: bone extraction     3D segmentation     computed tomography angiography (CTA)     region growth     medical image segmentation

1 基于改进的三维区域生长的骨骼组织外轮廓提取

1.1 改进的初始种子点选取方法

 图 1 头部影像灰度直方图Fig. 1 Histogram of head image in grayscale

1.2 改进的生长判定条件

 图 2 骨髓像素点示意图Fig. 2 Schematic diagram of marrow pixel points

2 改进的主动轮廓模型算法

2.1 改进的初始轮廓确定

2.2 改进的能量函数

Snake算法主要思想是通过定义一条初始的能量函数曲线,并将其初始位置定义在待分割物体的边界外围.通过求解该能量函数的极小值,能量函数曲线经过不断地演化,最终将收敛于目标区域的边界处,得到目标图像的轮廓.主动轮廓模型的定义为

3 算法实验与结果分析

 图 3 头部骨骼三维可视化效果图Fig. 3 Head skeleton 3D visualization figure

 图 4 不同的Snake算法对比图Fig. 4 Comparison between different Snake algorithms

 图 5 头部骨骼分割算法对比示意图Fig. 5 Schematic diagram of comparison between different head bone segmentation algorithms
4 结 论

1) 结合CTA影像的灰度特点,提出了三维区域生长的快速的骨骼区域种子点提取方法,并对三维区域生长算法的判定条件进行了适合CTA头部影像数据的合理修改.

2) 对于经典的Snake模型,本文结合医学影像的特征和功能对其进行了两个方面的改进:一方面是对Snake模型的初始轮廓确定的改进,这一改进减少了算法对于人工交互的要求;另一方面是在能量函数中引进了图像力这一参数项,使得该模型可以更好地解决当前的问题.

3) 最后给出实验结果并和传统的基于阈值和区域生长的骨骼组织分割算法的结果进行对比分析,证实本文提出的骨骼组织分割提取算法效果良好.

 [1] Witkin A P.Scale-space filtering[C]//Proceeding International Joint Conference Artificial Intelligence.San Francisco, CA:Morgan Kaufmann Publishers Inc., 1983, 2:1019-1022. [2] DeanBidgood W.Understanding and using DICOM, the data interchange standard for biomedical imaging[J].Journal of the American Medical Informatics Association, 2007, 4(3):199-202. Click to display the text [3] Freeman E, Robson E, Bates B, et al.Head first design patterns[M].Cambridge:O'Reilly Media, 2005:35-40. [4] Behrens T, Rohr K, Stiehl H S.Robust segmentation of tubular structures in 3-D medical images by parametric object detection and tracking[J].IEEE Transactions on Medical Imaging, 2012, 21(8):924-933. Click to display the text [5] Rinkel G J E, Djibuti M, Algra A, et al.Prevalence and risk of rupture of intracranial aneurysms:A systematic review[J].Stroke, 2008, 29(1):251-256. Click to display the text [6] Wink O, Niessen W J, Viergever M A.Fast delineation and visualization of vessels in 3-D angiographic images[J].IEEE Transactions on Medical Imaging, 2000, 19(4):337-346. Click to display the text [7] Udupa J K, Samarasekara S.Fuzzy connectedness and object definition:Theory, algorithms, and applications in image segmentation[J].IEEE Transactions on Graphical Models and Image Processing, 1996, 58(3):246-261. Click to display the text [8] Chan F T, Vese L.Active contours without edges[J].IEEE Transactions on Image Processing, 2001, 10(2):266-277. Click to display the text [9] Kass M, Witkin A.Snakes:Active contour models[J].International Journal of Computer Vision, 2007, 1(4):259-268. [10] Terzopoulous D, Fleischer K.Deformable models[J].The Visual Computer, 2008, 79(1):12-49. [11] Li P H, Zhang T W.Review on active contour model (snake model)[J].Journal of Software, 2000, 11(6):751-757. Click to display the text [12] Li J, Yang X, Shi P.A fast level set approach to image segmentation based on Mumford-Shah model[J].Journal of Computer, 2002, 25(11):1175-1183. Click to display the text [13] Cremers D, Kohlberer T, Schnorr C.Shape statistics in kernel space for variation image segmentation[J].Pattern Recognition, 2003, 36(9):1929-1943. Click to display the text [14] Stephen M P.Image processing and analysis at IPAG[J].IEEE Transaction on Medical Imaging, 2003, 22(12):1505-1518. Click to display the text [15] Pardo X M, Radeva P, Cabello D.Discriminated snakes for 3D reconstruction of anatomical organs[J].Medical Image Analysis, 2003, 7(3):293-310. Click to display the text [16] Montagnat J, Delingette H, Ayache N.A review of deformable surfaces:Topology, geometry and deformation[J].Image and Vision Computing, 2001, 19(14):1023-1040. Click to display the text [17] 李天庆, 张毅, 刘志, 等.Snake模型综述[J].计算机工程与应用, 2004, 40(11):43-45. Li T Q, Zhang Y, Liu Z, et al.A review on Snake models[J].Computer Engineering, 2004, 40(11):43-45(in Chinese). Cited By in Cnki (118) [18] 倪雅樱, 胡志忠.基于Snake模型的医学图像分割[J].上海生物医学工程, 2007, 28(3):165-170. Ni Y Y, Hu Z Z.Medical image segmentation based on snake model[J].Shanghai Journal of Biomedical Engineering, 2007, 28(3):165-170(in Chinese). Cited By in Cnki (5) [19] 王蒙, 戴亚平, 王庆林.一种新的FAST-Snake跟踪方法[J].自动化学报, 2014, 40(6):1108-1115. Wang M, Dai Y P, Wang Q L.A novel FAST-Snake object tracking approach[J].Acta Automatica Sinica, 2014, 40(6):1108-1115(in Chinese). Cited By in Cnki

文章信息

CAO Chunhong, AI Liang, XU Guangxing
CTA影像头部骨骼组织提取算法
Head bone tissue extraction algorithm based on CTA image

Journal of Beijing University of Aeronautics and Astronsutics, 2015, 41(6): 982-988.
http://dx.doi.org/10.13700/j.bh.1001-5965.2014.0502