Code for ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness

Add code button
avatar
Github Iconrgeirhos/Stylized-ImageNet

avatar
Github Iconmbuet2ner/local-global-features-cnn

avatar
Github Iconrgeirhos/texture-vs-shape

avatar
Github IconLiYingwei/ShapeTextureDebiasedTraining

avatar
Github Iconannstrange/breast-cancer-cnn

avatar
Github Iconjwyang/faster-rcnn.pytorch

avatar
Github IconBVLC/caffe/wiki/Model-Zoo

avatar
Github Iconopenimages/dataset/blob/master/READMEV2.md

avatar
Github Iconnaoto0804/pytorch-AdaIN