基于改进Canny算法的肺癌辅助诊断算法

Algorithm Simulation of Lung Cancer Auxiliary Diagnosis Technology Based on Improved Canny Algorithm

肺癌辅助诊断技术根本上是为了帮助医生进行诊断分析并提供一定的依据[1]。在边缘检测时运用了改进的Canny算法,在分割后进一步分割出肺结节。传统Canny算法采用的高斯滤波和高低双阈值不具有自适应性。使用者需要人为设定参数。采用自适应中值滤波代替高斯滤波,并用迭代法阈值选择算法得到双阈值,从而对Canny算法进行改进,并在肺部CT上进行试验,通过肺结节大小、位置、形状等生理特征判断是否患有肺癌。结果证明,上述方法可以更准确提取出肺结节轮廓并初步分析患者是否患有肺癌。

Lung cancer auxiliary diagnosis technology is to help doctors carry out diagnosis analysis and provide acertain basis. In this paper, an improved Canny algorithm is used in edge detection to segment lung nodules. The tra-ditional Canny algorithm adopts Gauss filter and high and low double thresholds which are not adaptive. The userneeds to set parameters manually. In this paper, an adaptive median filter was used instead of Gauss filter, and itera-tive threshold selection algorithm was used to get double thresholds, so Canny algorithm was improved, and lung CTwas used to test whether lung cancer exists by physiological characteristics such as the size, position and shape ofpulmonary nodules. The results show that this method can extract the lung nodule contour more accurately and analyzewhether the patient has lung cancer.