Automated Detection and Classification in Lung Cancer Screening
The system we are developing would work primarily to assist radiologist workflow by automating nodule detection. We are most interested in reviewing as many lung cancer screening CT's with nodules as we can, as this helps the computer learn. We will be using neural networks and deep learning computer models.
1. Develop software that can identify pulmonary nodules on lung cancer screening computed tomography with the accuracy that rivals or exceeds that of a practicing radiologist
2. Develop automated techniques to directly track nodules over time to detect subtle changes
Courosh Mehanian
Shawn McGuire