I was able to use the scaledMesh results from the mediapipe facemesh model along with the returned uv coordinates to overlay the video used for face detection onto the real-time 3d model of the detected face that I am visualizing in threejs. I am upgrading to use the face landmark detection model to be able to get the facial transformation matrices.
I do not see that the uv coordinates are returned with the results for the face landmark detection model. Is there a way to get these uv coordinates to map the texture onto the model created from the returned vertices?
Regarding the drawing of a 3d polygonal model from the data, I notice that the facemesh model returns 468 points and the face detection model appears to return 478 points. Will they use the same connectivity array to make the list of triangles to render a solid 3d mesh or is there a location I can get the list of triangle indices for the 478 points returned by the face landmark detection model? Is it possible that I need to run the models concurrently to be able to get both the uv coordinates along with the facial transformation matrices? Thank you.