Discrete Curvature

Replies to This Discussion

Permalink Reply by Will Pearson on September 11, 2015 at 3:38am

Hey Riccardo, nice work!

I thought I'd share some old code from a couple of years ago when I implemented "Taubin Tensors" for finding the principal curvature directions at each vertex of a mesh. It's Java unfortunately, but you'll find a link to the paper I referenced and my interpretation of the maths, so maybe this will help. No harm if it doesn't!

https://gist.github.com/pearswj/47989838b695b2553623#file-vertex-ja...

Permalink Reply by Riccardo La Magna on September 12, 2015 at 4:25pm

Hi Will,

thanks for sharing your code, it seems you've already done most of the work yourself. Is there any particular reason why this is not part of Plankton yet? I'd be glad to share my code, but I don't know whether it makes sense to push to Github what I've written so far.

Permalink Reply by Will Pearson on September 23, 2015 at 1:02am

Feel free to fork Plankton and add your additions. If you then open a pull request we can discuss further :)

Permalink Reply by Anders Holden Deleuran on September 11, 2015 at 5:12am

Terrific work Riccardo! There's also Mirco Becker's script. It operates on RhinoCommon meshes though, but it might help I imagine. It calculates the principal curvature directions using a rather elegant/terse method. I haven't tested it out too deeply, but we did give it a quick whirl for the AA Visiting School up Aarhus and it seems pretty legit.

Permalink Reply by Riccardo La Magna on September 12, 2015 at 4:21pm

Hey Anders! Thanks for pointing me to Mirco Becker's script. I like his method, the implementation is really simple and straightforward. But I've checked the results on the mesh above, and although the curvature distribution looks correct the curvature values seem to be pretty off. I think I'll push my code on Github and see what the bosses have to say :)

Permalink Reply by Daniel Piker on September 13, 2015 at 4:43am

Hi Riccardo, sounds great.

MeshMachine has a bit of code for calculating the cotan weighted mean curvature normal, based on that same paper. see here.

It would make a lot of sense though to integrate this into Plankton itself though, and Gaussian curvature as well would be nice.

Permalink Reply by Riccardo La Magna on September 14, 2015 at 9:42pm

Hi Daniel,

great, I didn't know MeshMachine already had some of the code implemented. I'll integrate my code with what you and Will have done so far and push it to Github once it's ready then.

Permalink Reply by Riccardo La Magna on February 13, 2016 at 8:51am

Hi everyone,

I've finally added the code for the computation of the principal directions. I've tried to push the changes to the Plankton repository but I'm getting an error which seems to be linked to permission limitations (Response status code does not indicate success: 403 (Forbidden)). Could that be the case? I'm pretty new to GitHub, so I might as well be doing something wrong.. Any help or hint much appreciated.

r

Permalink Reply by Daniel Piker on February 13, 2016 at 9:54am

Hi Riccardo,

Sounds great! This will be a really nice addition.

Did you try and submit a pull request?

Are you using the GitHub desktop app or something else?

Permalink Reply by Riccardo La Magna on February 13, 2016 at 11:06am

Hi Daniel,

I'm using the GitHub tools for Visual Studio. When I try to publish the branch I get the aforementioned error. If it's not related to permission restrictions I might be doing something wrong then..

Permalink Reply by Riccardo La Magna on February 14, 2016 at 5:02am

OK, I believe I've managed to successfully pull a request. If anything is wrong or missing please let me know. r