During my REU experience in the lab, I installed and tested various deep learning and neural network packages to see which packages would make the best learning tools, and then used this knowledge to create assignments for a new neural network course in the fall. Additionally, I used these packages to test alternative methods of gender recognition from the iris and co-author a research paper that was accepted in WACV 2017, included below.
Over the course of the school year, I designed and ran transfer learning experiments on detecting biometric spoof attacks which were featured in a textbook on deep learning and biometrics (to be published). I also conducted experiments using different network topologies to recognize genuine pupil reactions to positive and negative light stimuli.
