Early Lung Cancer Aided Detection Research Based On CT Image

Early detection, early remedy, early diagnosis is the key to improve the patient’s survival rate. But only 15% percent of lung cancer can be detected in the early time, the symptom is not obvious which lead to the result, what’s more patient can’t feel anything abnormal. So the key is to find early and diagnosis early. Early lung cancer computer aided detection has a great meaning.First, something corresponding to processing is introduced, which included denoising, enhancement and lung mask segmentation; Then, lung nodule segmentation,feature extraction, feature selection and classification has been recommended.Specifically, the main research details are as follows:1)Denoising, enhancement and lung segmentation’s algorithm and flow in preprocessing has been introduced which makes the fundament of next key part. Median filter were used in denosing, gaussian kernel based algorithm has been used in enhancement, lung segmentation proposed a threshold based frame.2)Lung nodule segmentation’s frame and algorithm, the equation of feature extraction or ideology are presented. Feature selection uses SFS technology. In order to classify positive and negative we use SVM classifier. At last, we analyzed and summarized the result of experiment.For the sake of verify the effect of the CADe we use the open dataset LIDC-IDRI to compare the result. The result indicate that the flow in the paper has a better validity.