基于边缘检测与支持向量机的关节软骨自动分割算法-第三军医大学学报.docVIP

基于边缘检测与支持向量机的关节软骨自动分割算法研究* 庞剑飞1，邱明国2，陈 伟3，刘 杰3，陈永林3 (400038 重庆，第三军医大学：学员旅十三队生物医学工程与影像学院医学信息与图像学教研室西南医院放射科摘要核磁共振成像MRI）图像的关节软骨自动分割法。利用像素特征构建二分类支持向量机（support vector machine，SVM）对股软骨、胫软骨及髌软骨进行分割股软骨、胫软骨及髌软骨Dice’s系数平均值分别为0.80、 0.76、0.74与分割具有较好的一致性结论 该算法能够准确、快速地分割出MRI图像中不同的软骨组织。 [关键词] MRI；软骨；分割；边缘检测；支持向量机[中图分类号]国家自然科学基金重大项F0205) [通作者] 邱明国，电话:（023 E-mail:qiumingguo@/kcms/detail/51.1095.R1548.002.html（2013-03-19） Segmenting articular cartilage automatically using edge detection and support vector machine Pang Jianfei1, Qiu Mingguo2, Chen Wei3,Liu Jie3, Chen Yonglin3 (1Thirteen Team of Cadet Brigade,2Department of Medical Informatics and Medical Image, College of Biomedical Engineering and Image,3Department of Radiology, Southwest Hospital, Third Military Medical University, Chongqing, 400038, China) [Abstract] Objective To explore the possibilities of automatic segmentation of articular cartilage in the MRI images. Methods We build three binary classifiers with edge-distance, intensity and other voxel features using support vector machine to segment the femoral cartilage, tibial cartilage and patellar cartilage separately. Firstly, we have designed an iterative procedure based on the feedback of the number of main edges to get an appropriate threshold for Canny operator and extract the main edges with it from MRI images. Secondly, it is to identify the different edges with some feature parameters, which can distinguish different cartilages synchronously. To improve the speed of segmentation, we have reduced the search space according to the cartilage edge and its anatomic location. Thirdly, we segment the cartilage tissues preliminarily with different classifiers which have been trained before. Finally, we use morphology operations to make those results better. Results The cartilage edge is smooth in automatic segmentation results and the average Dice similarity coefficient of femoral cartilage, tibial cartilage and patellar cartilage ar