Input dataset

[santech730]

I am starting to use your DeepCalib network (SingleNet) and struggling with the proper input images needed for the training. I was wondering if I could feed the network with the radial distortion parameters?

Thank you in advance.

Sanaz

[alexvbogdan]

Do you mean that you would use the Brown model with 4 distortion coefficients? In that case, you would need to update the code a little bit to work with 4-5 distortion parameters, and that is obviously a more difficult task for a network.