Follow
ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness. (arXiv:1811.12231v3 [cs.CV] UPDATED) http://arxiv.org/abs/1811.12231
ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness. (arXiv:1811.12231v3 [cs.CV] UPDATED) http://arxiv.org/abs/1811.12231