cvlab-epfl/LIFT

Spatial Transformer

Closed this issue · 4 comments

Hi,

Could you tell me how you use the spatial transformer function, ie which function is used to transform the input patch to the descriptor network according to the orientation?

kmyi commented

Hello,

This code does not have that part inside. The code for that and the training will be released (hopefully) soon. Drop an email address here, and I'll try not to forget to remind it to you when we get is ready.

Thanks!
Kwang

you will try not to forget to remind it to you when we get is ready.
lgdyangninghua@163.com

Thanks!

kmyi commented

Hello, I completely forgot, but the tf-lift implementation should resolve this issue. Also, the more recent lf-net also has everything implemented in tensorflow, and you can also train your own if you want.