基于磁共振图像的鼻腔边界分割及解剖结构提取  

MR image segmentation and extraction of anatomical structure in nasal cavity

作　　者：熊华晖[1] 李永[1] 李建红[2] 黄亚奇[1] 

机构地区：[1]首都医科大学生物医学工程学院,北京100069 [2]首都医科大学附属北京同仁医院医学影像中心,北京100730

出　　处：《生物医学工程与临床》2014年第1期1-4,共4页Biomedical Engineering and Clinical Medicine

基　　金：北京市优秀人才培养资助项目(2011D005018000009);国家科技支撑计划课题(2012BAI05B03);国家自然科学基金面上项目(81171422);北京市教育委员会科技计划重点项目(KZ201210025022)

摘　　要：目的探索基于系列磁共振图像获取鼻腔边界的有效分割方法,对图像进行准确分割并考察鼻腔的解剖结构。方法受试者为健康男性,年龄24岁。采用层间距为4 mm的冠状面头颈部MRI图像,用MATLAB编程对图像预处理。使用距离正则化水平集图像分割方法分割鼻腔的边界,并借助插值算法计算任意位置的鼻腔横截面积,以及通过对横截面积分布曲线的积分算出鼻腔体积。结果精确分割出鼻腔边界,获得其坐标,并可保存为图像格式或数据文件格式。给出左、右鼻腔横截面积随距离变化的连续分布曲线、鼻腔最小截面的位置,以及鼻腔体积。结论该方法可以有效地分割出几何形态复杂的鼻腔的边界并获得鼻腔的结构参数数值,所得到的边界坐标数据可方便地用于鼻腔结构的三维重建。Objective To explore an effective method of boundary segmentation on MR images of nasal cavity, and investigate the structure of the nasal anatomy. Methods A 24-year-old healthy male was enrolled, the coronal MR images of the neck and head with 4 mm slice spacing were performed and then preprocessed with MATLAB program. The distance regularized level set method were used to calculate the cross-sectional area of nasal cavity at any location by interpolation and the cavity volume was obtained by integrating distribution curve of cross-sectional area in nasal cavity. Results The boundary of nasal cavity on each image was segmented accurately, and its coordinates could be accessed in both graphic and numerical formats. The distribution curves of cross-sectional area in the left and right nasal cavities as function of distance and location of minimal cross-sectional area, as well as the total volume of nasal cavity were obtained. Conclusion It is demonstrated that the procedure could effectively extract the boundary of the nasal cavity on MR images and calculate the relevant structural parameters. The boundary coordinates extracted from the images could be used conveniently in the three-dimensional structural reconstruction of nasal cavity.

关 键 词：鼻腔 磁共振成像(MRI) 图像分割 水平集 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]