how can I train a single affine network?

microsoft / Recursive-Cascaded-Networks

[ICCV 2019] Recursive Cascaded Networks for Unsupervised Medical Image Registration

https://arxiv.org/abs/1907.12353

MIT License

361 stars 87 forks source link

how can I train a single affine network? #27

Closed oldriver-wang closed 4 years ago

oldriver-wang commented 4 years ago

I want to train an affine network, the inputs are src and tgt, but I don't know what the output should be. Flow field or W +b? And which loss should be taken there? cc loss or det_loss and ortho_loss? I am really confused.

zsyzzsoft commented 4 years ago

Try -n 0

oldriver-wang commented 4 years ago

thanks