Closed jeremmoore closed 1 year ago
I wonder if is it achieved by the model using blendshape.tflite and corresponding face mesh v2 model in the repository or just recording using mediapipe web demo?
just recording using mediapipe web demo
https://github.com/PINTO0309/facemesh_onnx_tensorrt https://github.com/PINTO0309/PINTO_model_zoo/tree/main/390_BlendshapeV2
I don't know what you are expecting.
Issue Type
Support
OS
Ubuntu, Other
OS architecture
aarch64
Programming Language
C++
Framework
TensorFlowLite
Model name and Weights/Checkpoints URL
https://github.com/PINTO0309/facemesh_onnx_tensorrt/issues/4
Description
A demo video is attached in readme. md. I wonder if is it achieved by the model using blendshape.tflite and corresponding face mesh v2 model in the repository or just recording using mediapipe web demo? For now, I cannot download the corresponding face mesh v2 model mentioned in face blendshape model card which outputs 478 landmarks. Only face mesh with 468 landmarks is what I can find now.
Relevant Log Output
No response
URL or source code for simple inference testing code
No response