Brummi / MonoRec

Official implementation of the paper: MonoRec: Semi-Supervised Dense Reconstruction in Dynamic Environments from a Single Moving Camera (CVPR 2021)

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

How to train with only Monocular data sequence?

Owen-Liuyuxuan opened this issue · comments

As I was reading the paper, it seems that static stereo plays quite a role in the training process?

Hi,
thank you for being interested in the project!
Yes, during training we require stereo frames, otherwise there would be no training signal for moving objects.
Inference runs only on mono frames.
Best,
Felix

@Brummi What does the "training signal" for moving objects indicate ?