Head Tilt Angle Classification using CNNs

This projects employs a Neural Network to perform Head Tilt Angle Classification, acheiving a top accuracy of 81% to detect faces facing downward, straight ahead or upward. It was done in collaboration by Sacheth Sathyanarayanan, Hari Ramakrishnan and Ish Kaul as the final project to Princeton University's COS 485 (Neural Networks) class. Please read through Report.pdf file for details.