CT图像肺结节自动检测

肺癌是对人类生命健康危害最大的恶性肿瘤之一。计算机辅助诊断系统对肺部CT图像进行自动分析后,可提示医生可疑肺结节,从而克服医生在诊断中的一些主观因素,为此本文提出了一种基于胸部CT图像的可疑肺结节自动检测算法。首先,根据胸部组织的特殊结构,利用一种新的分割算法提取出肺实质部分;在此基础上提取出灰度与结节相近的感兴趣区域,包括结节、肺血管、支气管;然后,以已标记的结节数据作为样本集,计算结节的面积、灰度均值、灰度方差、圆形度、形状矩、体积、球形度等特征值,利用最近邻法建立分类器判别函数;最后,计算测试集感兴趣区域的上述特征,对其进行判别、分类,并标记出结节。试验结果表明,该算法综合考虑了肺结节特征,具有较高的准确度。

Automated Detection of Lung Nodules Based on CT Images

Lung cancer is one of the most malignant tumour in human life. The computer aided diagnosis system can automatically analyze the lung image and prompt lung nodule to doctors, thereby overcoming some subjective factors from doctors in diagnosis. This paper presents an automatic detection method for lung nodules based on chest CT images. First, according to the special chest structure, a new segmentation algorithm was proposed to extract the lung parenchyma. The regions of interest （ROI） where gray was likely as nodule, such as nodules, pulmonary vessels, airways, were extracted. Second, taking marked nodules as sample set, the area, average gray, variance, shape matrix, volume, spherical degree were set as discriminating features, the discrimination of classifier was devised based on the nearest neighbor algorithm. Last, such features of ROI in testing set were calculated and classified. Experimental result showed that the algorithm considered the characters of lung nodule generally and achieve high accuracy.