Creating and testing Nvidia's End-to-End CNN model agains traditional computer vision practices to compare performace differences and generalisation of the models.
Demonstration of the model in practice can be found right here: https://www.youtube.com/watch?v=DaUzsRCO_1M
The Deep Learning based on the Nvidia's End-to-End CNN model performed the best out of the three algorithm's tested achieving a 96% accuracy score on validation set. The models learned to generalised enough to drive on roads without significant lane markings as well.