基于SVM算法的乳腺X片微钙化点自动检测技术.pdfVIP

_______________________________________________________________________________ 基于 SVM 算法的乳腺 X 片微钙化点自动检测技术 万柏坤，王瑞平，朱欣，赵丽，刘庆凯，李树楠 （天津大学 精密仪器与光电子工程学院 生物医学工程与科学仪器系,天津 300072） 摘要 支持矢量机(SVM)是一种新的统计学习方法，其学习原则是使结构风险最小，而非经 典学习方法所遵循经验风险最小原则。这使得 SVM 具有更强的泛化能力。并且，由于 SVM 求解的是凸二次优化问题，使之能保证所找到的极值解就是全局最优解。本文首次将 SVM 算法用于乳腺 X 影象微钙化点自动检测中，对临床实际病例的试用结果表明，同目前常用 的基于经验风险最小的人工神经网络(ANN)分类方法相比，SVM 具有更高的识别率，值得应 用推广。 关键词 小波变换支持矢量机，结构风险最小，经验风险最小，微钙化点，乳腺 X 片 4、（英文）题名、作者姓名及所在单位、摘要、关键词： Automatic Detecting Technique of Micro-calcifications in Mammogram Based on Support Vector Machine WAN Bai-kun, WANG Rui-ping, ZHU Xin, ZHAO Li, LIU Qing-kai, LI Shu-nan (Department of Biomedical engineering Science Instrument, College of Precision Instrument Opto-electronics in Tianjin University, Tianjin 300072, China) Abstract: Support vector machine (SVM) is a new statistical learning method. Compared with the classical machine learning methods, its learning discipline is to minimize the structural risk instead of empirical risk the learning discipline of classical methods, and it gives better generative performance. Because SVM algorithm is a convex quadratic optimization problem, the local optimal solution is certainly the global optimal one. In this paper applies SVM algorithm is applied to detect the micro-calcifications in mammogram at the first time. The algorithm is tested with 10 mammograms and results show that the algorithm achieves a higher true positive in comparison with artificial neural network (ANN) based on the empirical risk minimization. Key words: Support vector machine (SVM); Structural risk minimizatio