改进的Canny算子与SIFT算法相结合的医学图像特征提取  

Improved Canny Operator and SIFT Algorithm Combining Medical Image Feature Extraction

作　　者：刘桥[1] 方茜[1] 程欢[1] 杨正坤[1] 

机构地区：[1]长沙理工大学电气与信息工程学院,湖南长沙410114

出　　处：《工业控制计算机》2012年第12期63-65,共3页Industrial Control Computer

摘　　要：微创外科手术中的图像特征快速提取与分析,可以使计算机具备图像实时识别的能力,提高手术的成功率。对常用的边缘与角点特征进行了分析,提出了边缘与角点相结合的特征提取方法。以二尖瓣索修复微创手术图像为例,对常用的几种边缘算法和改进的Canny算法进行了分析比较,利用改进的Canny算子所得到的较为完整的检测边缘信息,运用SIFT算法提取图像角点,剔除了一些伪特征点,得到了图像匹配时所需要的更加精确的信息。为后续的图像融合与复原以及三维图像重建打下了良好的基础。The rapid extraction and analysis of Minimally invasive surgery image features,can make the computer have image re- al-time recognition ability,and improve the success rate of surgery.The common edge and corner point features,this paper analyzes the edge and corner combining feature extraction method,With Mitral Valve Chordae Repair Minimally invasive surgery image as an example,the commonly used several kinds edge algorithm and improved Canny algorithm are ana- lyzed and compared,and the use of improved Canny operator the relatively complete detection edge information.Using the SIFT algorithm,images of angular point,eliminate some of the pseudo feature points,obtained the image rnatching needed for more accurate information,

关 键 词：特征提取 边缘检测 角点检测 SIFT CANNY 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]