Screenshot (45)Lack of classifier robustness to CT acquisition parameter variations is a barrier to widespread adoption of CAD systems. As part of his PhD research, CVIB graduate Dr Danny Chong developed a new Robustness-Driven Feature Selection (RDFS) algorithm that preferentially selects features that are relatively invariant CT technical factors.

Working with CVIB faculty and staff, Dr. Chong showed that use of RDFS with 3D texture features improved CT classification of fibrotic interstitial lung disease in multi-center clinical trials. The research was published in the January 2016 issue of the IEEE Transactions on Medical Imaging.