中文摘要：

图像处理在计算机辅助医疗诊断系统中有着广泛的应用,本文在水平集方法的基础上,提出了一种肿瘤图像识别方法。首先,采用改进的水平集方法对肿瘤图像进行分割,获得肿瘤轮廓特征;其次,提取肿瘤的轮廓特征参数,对贝叶斯分类器进行训练,然后用训练的分类集进行肿瘤良恶性分类识别。通过对CT图像的肝脏肿瘤进行良恶性诊断,显示所开发的计算机辅助诊断系统是有效的。

英文摘要：

Image processing is widely used in computer aided medical diagnosis system.In this paper,a method based on level set method is proposed for tumor images recognition.Firstly,an improved level set method is used to segment tumors from CT images and extract the contours of the tumors;secondly,the contour feature parameters are extracted from the contours.Then the NaiveBayesian-based method classifies the tumors into benign and malignant ones.The experiments for diagnosis of liver tumor from CT images show that the computer aided diagnostic system is effective.