How to Predict 3D Faces Using Models?

[google]

How to Predict 3D Faces Using Models?

[zhouhao-learning]

@kartynnik
Thank you for sharing. I'm still looking for ways to implement Python

[mgyong]

The 3D face model is significantly more difficult than the facedetector_front.tflite we have in the model folder. We recommend that you read the publication for Face Mesh

What is the use case you are looking at for 3D face detection?

[zhouhao-learning]

@mgyong
Sorry, maybe the meaning of the expression is wrong. I actually want to use the model you provided to do face detection. I want to get the following detection effect:

But I don't know how to use your model and build prediction code. Can you help me?

[mgyong]

@zhouhao-learning Have you gone through the Face detection for Android running GPU example?
If you have issues with understanding some of the Android APIs for MediaPipe, you can start with simpler example

[zhouhao-learning]

No, I didn't run the Face detection for Android running GPU example. I don't want to run on Android right now. I just want to write a predictive code that can achieve the face alignment effect, but I don't know how to build the prediction code. Can you help me? Thank you.

[mgyong]

@zhouhao-learning This 3D face mesh issue is out of scope for project currently

[zhouhao-learning]

@mgyong
No, if it is beyond the scope of the project, how do you achieve the following effects?

[kartynnik]

@zhouhao-learning If you are interested in getting the effects by any means, you might find it useful to know that this 3D face mesh technology is exposed via the Augmented Faces API in ARCore.