1 paper
Zhendong Zhang, Cheolkon Jung, Xiaolong Liang
Recent works show that deep neural networks trained on image classification dataset bias towards textures. Those models are easily fooled by applying small high-frequency perturbat…