A question about paper, thanks your help.

[Rythsman]

Hi @AlessioTonioni, when I read your paper, I can't understand where you use unsupervised adaptation.
In your paper, SGM/AD-CENSUS + CCNN are used as ground truth labels(in your paper you write 'CCCN'), but when i read the CCNN papar, it just use the kitti's ground truth to calculate the confidence, so can you tell me how to deal with datasets that don't have ground truth?
Thx.

[AlessioTonioni]

We have not retrained CCNN for our paper but simply used the pre-trained one provided by the authors available here, so no ground truth is used for the adaptation process.

[Rythsman]

thx.