huynhlam / dav Goto Github PK

Will the code be released?

Hi,

Since the resolution of RGB frames and depth are different, is any alignment operation required? Or I just need resize both RGB image and depth to the same resolution?

Thanks.

Hi, I want to project the 3D points features back to an 3D tensor of size CHW, Is there a fast method to do it?

Will the code be open source ?

Hello, I see the description From your paper“ ScanNet dataset comprises of 2.5 million RGB-D images acquired from 1517 scenes. For this dataset, we use the training subset of 20K images provided by the Robust Vision Challenge 2018 [9] (ROB). Unfortunately, the ROB test set is not available, so we report the results on the Scannet official test set of 5310 images instead."

Could your tell me how to get the lists :"training subset of 20K images" and " test set of 5310 images "?

I am very appreciate for your reply!

best regards!