Why do you use rollaxis to turn data array into channel_first and then reshape again to channel_last?

0bserver07 / One-Hundred-Layers-Tiramisu

Keras Implementation of The One Hundred Layers Tiramisu: Fully Convolutional DenseNets for Semantic Segmentation by (Simon Jégou, Michal Drozdzal, David Vazquez, Adriana Romero, Yoshua Bengio)

https://arxiv.org/abs/1611.09326

MIT License

197 stars 54 forks source link

Why do you use rollaxis to turn data array into channel_first and then reshape again to channel_last? #13

Open MuyanXiao opened 6 years ago

MuyanXiao commented 6 years ago

The images are changed during this reshape.