Patrik Huber

Hi Pranav, Thank you for reporting this. Looks like the default template argument of `clip_polygon_to_plane_in_4d` is specified twice and AppleClang doesn't like that, and for some reasons other compilers accept...

Thanks a lot for reporting back. Great that this worked. I'll re-open / leave this open, as I should probably check this and fix it / change it permanently.

Hi @DenysVolkov, eos does not need libjpeg, I think this comes as a transitive dependency from OpenCV. Maybe the eos CMake scripts have to be updated in the way they...

Regarding the mouth, that's probably because some of the blendshape coefficients (e.g. the one(s) that open the mouth?) are not exactly zero after the fitting. You could have a look...

I'm not sure what you mean, can you elaborate? You can upload files to GitHub issue.

Definitely having the ability to add a full head or transfer the fitted face to a full head would be a very welcome addition! This has also been discussed before,...

I don't think I've encountered that before. It looks like it could either be a rounding/1-pixel inaccuracy in the renderer, or something with the texturing/texture extraction. I wonder where it's...

Ok, interesting. Can you provide the input landmarks and image, and which fitting exactly you ran and with which parameters? Does it occur with every image/rendering, or only with a...

@chrisluu I don't have any OpenGL code. But it should be easy to do, as all matrices should be OpenGL compatible, and glm works well with OpenGL.

My observation is that it takes up to 50 or so iterations for the parameters to fully converge. However, visually, there is usually no difference after 5 or so iterations....