摘要
肺结节的智能识别对肺癌的诊断至关重要。为了在大量的肺部CT图片中准确智能识别肺结节,我们研究了一个基于规则及多特征跟踪的肺结节计算机辅助检测方法。其中,采用活动轮廓模型的分割方法实现候选肺结节分割,采用基于规则的决策方法以及多特征跟踪方法实现肺结节分类。实验证明,该肺结节的智能检测方法满足肺结节计算机辅助诊断的要求。
Intelligent recognition of lung nodules which are distinct signs of lung cancer is important for diagnosis of lung cancer.In order to accurately identify lung nodules in a large amount of thorax CT images,a method of Computer-Aided Detection（MCAD） for lung nodules based on rules and multi-feature tracking was proposed in this paper.In this proposed method,candidate pulmonary nodules were segmented by GVF snake model;lung nodules were classified based on rules and multi-feature tracking method.Experiments demonstrate that the MCAD provide a good means for computer-aided diagnosis of lung nodules and cancer.
出处
《生物医学工程研究》
2010年第2期79-83,105,共6页
Journal Of Biomedical Engineering Research
基金
中国博士后基金资助项目(20090450866)
广东省自然科学基金资助项目(8451064101000631)
广州市番禺区科技攻关项目(2009-Z-108-1)
教育部博士点基金资助项目(200805610018)
广东省教育部产学研结合项目(2009B090300057)
关键词
肺结节
计算机辅助检测
分类
医学图像
识别
Lung nodules
Computer-aided detection
Classification
Medical image
Recognition
