microsoft / scenelandmarklocalization Goto Github PK

Where can I get depth images of the released dataset. I want to assess the dataset by scene coordinate regression methods.

Hello,
do you have any plans to publish 3D model files of your datasets？
If possible, I really hope to get your model data for academic research.
I am looking forward to your reply！

Are these functions defined in conv2d_layers.py?

Dear Authors,

When is the intended date for code release?

Thanks.

hey,I'm really interested in your work.

Therefore, I want to experiment on other scenes, but I have not found the program code about how to select the representative landmark in the scene. Can you release the code about how to select the representative landmark in the three-dimensional scene?

I am looking forward to your reply!

Hello. In the related paper, a neural bearing estimator is proposed, but I cannot find that network in this source code. Is that structure deprecated?

What are the scale unit of the 3x4 pose in each pose.txt?
is it meter? (paper reports cm but seems like too small)

Thanks for your great work. I want to reproduce the results on 7 scenes dataset in the paper but it seems that there are only configs for indoor-6 datasets in current code. How can I run it on 7 scenes dataset?

Hi, a great job！
May I ask when to release the source code？
Not in November 2022？