Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Question about generating WTN's ground truth #3

Open
artvandelay opened this issue Sep 21, 2017 · 5 comments
Open

Question about generating WTN's ground truth #3

artvandelay opened this issue Sep 21, 2017 · 5 comments

Comments

@artvandelay
Copy link

It would be great if you could share the code (or talk about) how to generate the ground truth for WTN. More specifically, what does bin 0 correspond to? In the paper it says

In this case, bin 0 corresponds to background or regions within 2 pixels of an instance boundary...

Does this just mean all the pixels outside of the instance are in bin 0? or bin 0 has all the pixels outside the polygon and 2 pixels inside it?

Thanks!

@amosella
Copy link

Hi,

Maybe I can help you. Right now the code in order to generate the ground truth is in the repository, you can find this code here https://github.com/min2209/dwt/tree/master/matlab. Answering your question, far as I understood the bin 0 contain the background and the pixels at a distance of two pixels of the instance boundary.

@min2209
Copy link
Owner

min2209 commented Sep 25, 2017

Yes, amosella is right.

@jigarCrowdAI
Copy link

Thanks for your comment.

I looked at the code, in generate_GT_cityscapes_unified at line 12,
for i=1:length(depth_bins)-1 depth_map(depth_map > depth_bins(i) & depth_map <= depth_bins(i+1)) = i-1; end

Where the depth_bin is the boundaries list like this
boundaries = [0,1,2,3,4,5,7,9,12,15,19,24,30,37,45,54,Inf];

This doesn't seem to confirm that all pix at distance 2 ( i.e. all values <= 2 from bwdist) should be part of bin 0. In this case it seems like it's only for <=1.

What am I missing? Thanks again

@min2209
Copy link
Owner

min2209 commented Sep 25, 2017

Ah, yes, you're right about that - however, keep in mind that the code operates on a 0.5x downsampled image (2048x1024 -> 1024x512), so that's the discrepancy.

@jigarCrowdAI
Copy link

Ah, I see. Now it makes sense! Thanks

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

4 participants