1 paper
Robert Geirhos, Carlos R. Medina Temme, Jonas Rauber +3
We compare the robustness of humans and current convolutional deep neural networks (DNNs) on object recognition under twelve different types of image degradations. First, using thr…