改进的3D Canny算子在MRI乳腺图像分析中的应用  

Application of Improved 3D Canny Operator for MRI Breast Image Analysis

作　　者：杨娉娉 武林会 米红妹 朱建明 

机构地区：[1]中国计量大学信息工程学院与生物医学工程研究所,杭州310018

出　　处：《计算机测量与控制》2017年第9期143-145,149,共4页Computer Measurement &Control

摘　　要：为了突破Canny算子普遍适用于二维图像的局限性和实现大批量三维MRI乳腺图像的精确边缘提取,结合了基于二维Canny算子和三维边缘检测方面的相关知识,提出了一种基于二维非极大值理论的三维非极大值抑制新方法和迭代法自适应双阈值选取的方法;现有的三维边缘检测算子在非极大值抑制方面选取像素点,考虑x,y,z三个轴的信息,通过比较该像素点与梯度方向上两个点的幅值大小;为了简化模型,在三维空间内寻找一个角,用它取代与3个轴的梯度方向角,在此基础上实现线性插值;根据图像二维直方图的形状,采取迭代法双阈值的方法,避免了传统的人工选取阈值的繁琐;通过主观评价对边缘检测结果进行分析,结果表明,改进的3D Canny应用在MRI乳腺图像中也能达到比较好的结果。Canny operator has been widely used in two-dimensional image analysis. In order to realize its application in edge extraction for 3D MRI breast images, an adaptive, double-threshold selection method, based on 2D non-maxima suppression technique, was pro- posed. Current three-dimensional edge detection operators use the information from the x, y, and z axes to select the image pixels by non- maxim suppression. By comparing the amplitudes of the pixel with two other pixels along the gradient direction. In order to simply the model and improve the accuracy, a corner in the three-dimensional space was searched to replace the corner along the gradient direction of x, y, and z axes, thus linear interpolation was achieved. Based on the shapes of the two-dimensional histogram, two-threshold iteration method was used to replace manual threshold selection. Edge detection results were analyzed by taking subjective evaluation . Our results indicated that, by applying improved 3D Canny operator, breast MRI image edge detection also can get a good result.

关 键 词：边缘检测 非极大值抑制 迭代法 乳腺图像 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]