Lung cancer is one of the malignant tumor threatening human health in the world, in many parts of the world lung cancer is a major cause of cancer deaths.Early detection and treatment is very important to improve the survival rate of patients with lung cancer, so how to early det-ect and diagnose is the research focus and hot spot of the medical profession. And early forms of lung cancer is lung nodules, so it is very important to correct detection of lung nodules.Ho-wever, the detection of lung nodules relies mainly on the doctor's experience, has a great deal of subjectivity, and, with the increment of the doctor read, they prone to misdiagnosis phenom-enon.This paper, by using 19 characteristic parameters that are the gray average, gray variance, energy, contrast, entropy, inverse difference moment in four directions as well as local binary patterns,uses support vector machine (SVM) for training, testing, and then the corresponding classifier is obtained.In this paper, the average accuracy is 84.33%. |