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

issue with texture stitching #7

Open
SriGanesh130 opened this issue Jul 16, 2019 · 18 comments
Open

issue with texture stitching #7

SriGanesh130 opened this issue Jul 16, 2019 · 18 comments

Comments

@SriGanesh130
Copy link

created a .ply file from obj file using meshlab.Ran TextureStitching with mesh and composite texture and mask of that texture. it saying:
[EXCEPTION] InitializeGridChartsActiveNodes: Node already covered

Tried with the partial textures too, its giving core dumped. Give me some suggestion about what is texture format specifier and confidence format specifier

@mkazhdan
Copy link
Owner

Would you be able to share your files?

@SriGanesh130
Copy link
Author

data contains input images, part unwraps, segmented mask ::
data.zip

@mkazhdan
Copy link
Owner

Can you send the .ply file as well?

@mkazhdan
Copy link
Owner

Also, do you have a composite texture image?

@SriGanesh130
Copy link
Author

This is perfectly stitched texture, If I want to generate the composite texture from part textures from your algorithm, How to generate with high accuracy.
data1.zip

@mkazhdan
Copy link
Owner

To use the algorithm, you need to provide three things:

  1. a textured mesh,
  2. a composite texture file, and
  3. a mask file.
    Given the consistent texture maps in the folder "unwraps", you can get a composite by iterating over the texels in the texture image and using the texel from the best "unwrap" image. A simple way to define "best" is to require that the texel be visible and, of all visible texels, it has a surface normal that is pointing most directly towards the camera. (e.g. it maximizes the dot-product between the normal and the negative camera direction).

Independently, you are getting a "Node already covered" error because the texture mapping in result.ply is not injective. In particular, triangles 1502 and 8389 map to the same triangle in texture space.

@nikanfds
Copy link

nikanfds commented Feb 3, 2020

I'm having the same issue. and in my mesh there are patches that map to the same triangle on texture, therefore I need them to be the same color in 3D space. Is that possible?

@mkazhdan
Copy link
Owner

mkazhdan commented Feb 3, 2020

I'm afraid it's not supported at the moment.
The current assumption is that independent patches are free to have colors assigned to them independently.

@nikanfds
Copy link

nikanfds commented Feb 4, 2020

if I remove the duplicated patches (the ones correspond to the same triangle on 1024*1024 texture), therefore the mesh would have some holes; but then I can run it through your code and get the final processed texture, right?
Then at the end I can put back the removed patches and give them the color based on the output of "TextureSignalProcessing"

@mkazhdan
Copy link
Owner

mkazhdan commented Feb 5, 2020 via email

@NengQian
Copy link

I have the same error message too.
Is that no injective texture mapping the only reason for this error?
Are there any methods to determine if my mesh and texture mapping is injective?
If they are not injective, how to find which triangle face actually share the same uv?
Thank you!

@mkazhdan
Copy link
Owner

Can you share your data?

@mkazhdan
Copy link
Owner

Also, can you try the newly posted code/executables? It should give you information about the two triangles causing trouble.

@NengQian
Copy link

NengQian commented Oct 9, 2020

@mkazhdan
Thanks for your reply.
The data is regarding our new project so I guess I cannot share the data now.
I am using the TSP.x64.zip file in the repo, it was updated 17 month ago. Is it the newly posted one?
It did not give me the information about the two triangles causing the trouble...

@mkazhdan
Copy link
Owner

mkazhdan commented Oct 9, 2020

It should be. When I do an ls -l in the directory I see a time-stamp from 9/29/20.

@NengQian
Copy link

I rechecked the date of the TSP.x64.zip file in the repo, it was updated on 2018.
I guess this 9/29/20 could be the date your decompressed the executable file in your local computer?

I reran the TextureStitching.exe from that zip file, and it only gave me the error message:
[EXCEPTION] InitializeGridChartsActiveNodes: Node already covered

Where can I find the newly posted one?

@mkazhdan
Copy link
Owner

I apologize. That was an old link to the executables. You should be able to grab the new executables from:
http://www.cs.jhu.edu/~misha/Code/TextureSignalProcessing/TSP.x64.zip

@NengQian
Copy link

Thanks!
Now it shows the error message properly with the triangle Ids.

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