Pritish Samant

Evaluating the robustness of an autonomous vehicle's object recognition system against adversarial attacks using FGSM. (Thesis) 1. Training Deep Learning model CNN (convolutional neural networks) for object detection with TensorFlow & Keras. 2. Classifications of traffic sign images using CNN and then generated adversarial images using FGSM (Fast Gradient Sign Method). 3. Implemented Adversarial Training to mitigate the misclassification of traffic sign images and robu