Computer simulation for segmentation of lung nodules in CT images

Automated lung nodule detection through computed tomography (CT) image segmentation is a new and exciting research area of medical image processing. We are currently developing a nodule detection system. For the testing stage we have developed a method to insert simulated lung nodules into CT images. The simulated nodules can be used to produce corner cases to provide a better test environment for the segmentation technique than would be available through clinical data. The synthetic lung nodules produced by this program are based on a 2D Gaussian structure. This is modeled on the study of the structure of real lung nodules. We have also developed a lung segmentation technique, which is the first stage of our nodule detection system. The lungs are segmented using a combination of thresholding, morphology, 3D region growing, and volume analysis