Can I use my own head mesh and migrate my own head mesh into MetaHuman?

Now I have a face mesh which is obtained from scanning and registration. Is it possible to use my head mesh in metahuman? I would like to use my own head mesh while still use the skeleton, weights, etc. in order to use the expression system of MetaHuman.

Is it possible?

What I have tried:
I exported the head of metahuman as FBX. I rigidly aligned my own mesh to the metahuman head using Warp3. I transfered bones and weights to my own mesh using Blender. Finally the generated FBX was imported into UE4 and I connected Amination, BluePrint, etc. to the new head. But the head does not move with Live Link. Something must be wrong in this process…

I believe you also have to add the live link BP to the body as well because the head movement is driven from there and not the face mesh.

Thank you! Actually I only need the expression system. Do you mean that driving facial expression relies on the body? Is it possible that I drive the face mesh only regardless of the head roation?

The head rotation comes from the body, I’m using faceware with live link, if I only add the blueprint to the face mesh for the animation mode it doesn’t rotate the head, if I add the faceware blueprint to the body animation as well it then drives the head as well.

Now you can! We just release the MetaHuman plugin for Unreal Engine 5 along with the new Mesh to MetaHuman Tool. Check it out: Using Mesh to MetaHuman in UE | Unreal Engine - YouTube

Thanks I just saw the update - incredible!

I had one question - what programs/equipment do I need to use to scan a face that can be used in this way?

Thanks again

Can I ask what programs/equipment you used to make a face scan?

Thanks

We have some cameras and we can construct point cloud for the scanned face. Since this is not done by me, I am not sure which software my collegue used. But there exist many available ones.

For the mesh, we use Warp4D to wrap the mesh to the scanned point cloud.

1 Like

Thank you for the reply! Appreciate it!

By the way, now I learned to use RealityCapture to do multi-view construction of the point cloud and the initial mesh of the scanned face.