Design And Implementation Of Key Algorithm On Pulmonary Nodules Aided Detection

Lung cancer has become the most major threat to human health that over one million people all over the world die from it ever year. In order to reduce the mortality, it is necessary to diagnose and cure lung cancer earlier as much as possible. And with the development of Image Processing, Pattern Recognition, Scientific Calculation and Visualization, Computer Aided Diagnose (CAD) technology has been valued gradually. Detecting pulmonary nodules in CT image by CAD is an effective way to diagnose lung cancer earlier.Therefore, we did some research about pulmonary nodule detection in CAD, designed and implemented some key algorithm on pulmonary nodule aided detection. Firstly, we preprocessed the 3D CT images by using interpolation algorithm and pulmonary parenchyma segmentation algorithm, making detection area in pulmonary parenchyma. Secondly, we segmented lung vessels by modified C-V model segmentation algorithm and got rid of them from pulmonary parenchyma, reducing the interference of lung vessels and detection area. Thirdly, for suspected nodules detection, we did threshold limitation algorithm as well as Hessian Matrix multi-scare detection algorithm for compare. At last, We analysed the features of nodules, then distinguished the suspected nodules by the Fisher linear classifier. Finally, the positive nodules were extracted.Our experiments show that the pulmonary nodule detection method we designed can detect pulmonary nodules effectively especially for solitary pulmonary nodules. Therefore, the method can help improve the efficiency and accuracy of diagnosis.