facebookresearch / meshtalk

Code for MeshTalk: 3D Face Animation from Speech using Cross-Modality Disentanglement
Other
364 stars 56 forks source link

Export to fbx / csv animation #14

Closed fmac2000 closed 2 years ago

fmac2000 commented 2 years ago

As title suggests, what would be the approach to converting the tensor geometry to a animation file instead of rendering a video of the animation?

fmac2000 commented 2 years ago

Fantastic work by the way guys!

alexanderrichard commented 2 years ago

In the animation script in this line https://github.com/facebookresearch/meshtalk/blob/main/animate_face.py#L97 the result tensor contains the 3D vertex positions for all vertices of the face mesh and for every time step of the animation. Add the definition of the faces fromhttps://github.com/facebookresearch/meshtalk/blob/main/assets/face_template.obj to each frame and save in in wavefront or .obj format and you can load it into any animation software like blender.

fmac2000 commented 2 years ago

Awesome, thanks for the fast response I'll get to work 👯‍♂️