应用多尺度三维图搜索的SD-OCT图像层分割方法  

SD-OCT Image Layer Segmentation Using Multi-scale 3-D Graph Search Method

作　　者：牛四杰[1] 陈强[1] 陆圣陶 沈宏烈 

机构地区：[1]南京理工大学计算机科学与工程学院,南京210094

出　　处：《计算机科学》2015年第9期272-277,共6页Computer Science

基　　金：青蓝工程(D040201);中央高校基本科研业务费专项资金(30920140111004)资助

摘　　要：频谱域光学相干层析技术(SD-OCT)是一种广泛应用于眼科领域的成像技术,视网膜组织层分割对视网膜疾病诊断起着至关重要的作用。传统的三维图搜索方法能够同时分割k(k≥1)个三维面,但其存在时间复杂度高、分割病变图像鲁棒性弱等问题。在传统三维图搜索模型的基础上引入多尺度思想,提出应用多尺度三维图搜索的SDOCT视网膜图像分割方法。首先根据每个组织层的特点,为每层构造一个合理的顶点权重;然后利用相邻列的最大与最小高度差构造列约束限制,改进表面的平滑约束条件;最后利用低尺度的图像,应用三维图搜索方法进行粗分割,逐步向高一尺度应用三维图搜索方法进行单表面细分割。使用改进算法对3组正常眼睛及1组老年黄斑变性视网膜图像进行分割,并将结果与手动分割及传统三维图搜索方法进行比较,实验结果表明,改进算法能够准确有效地分割出3个层边界(边界位置绝对误差是3.86±2.50μm),并且接近于手动分割结果(3.78±2.76μm),优于传统三维图搜索方法(7.92±3.31μm)。Spectral domain optical coherence tomography （SD-OCT） imaging technique is widely used in the field of ophthalmology. The segmentation of retinal tissue layers plays a vital role in the diagnosis of retinal disease. The tradi- tional 3-D graph search method is able to segment k surfaces simultaneously. Problems of this algorithm associated with high time complexity and weak robustness of segmenting pathological retinal images have limited its utility. Multi-scale theory was introduced to the traditional 3-D graph search modal, and a 3-D graph search algorithm based on multi-scale for segmenting retinal images was proposed in this paper. Firstly, reasonable cost functions are constructed for each sur- face according to the characteristics of the layer. Then the minimum and maximum differences in height of adjacent col- umns are utilized to construct inter-column constraints to improve smoothness constraints of surfaces. Finally, 3-D graph search method is used for coarse segmentation of lower-scale images and then a search region of higher-scale ima- ges is redefined for a more accurate result. The improved algorithm was evaluated on 3 groups of normal eyes and 1 group of abnormal eye with age-related macular degeneration. The experimental results were compared with manual seg- mentation and traditional 3-D graph search method. The results demonstrate that the improved method can effectively detect 3 layer surfaces （the mean absolute boundary positioning difference is 3.86 ±2.50μm） more closely to manual segmentation （3.78±2.76μm）, and it is better than traditional 3-I3 graph search method （7. 92±3.31μm）. Keywords Spectral domain optical coherence tomography, Retinal image segmentation, Multi-scale, 3-D graph search

关 键 词：SD-OCT技术 视网膜图像分割 多尺度 三维图搜索 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]