your code is very cool!Few people have found this treasure of yours。
You have a 26-dimensional expression coefficient estimate。
I have an idea：
A person had two pictures, one with no expression, the other with exaggerated expression; Based on the 3D mesh estimated by the picture without expression, the expression coefficient of another picture and the final 3D mesh were directly estimated. In this way, we can better estimate the person's 3D facial mesh in real time.
thanks for your code。
your code is very cool!Few people have found this treasure of yours。 You have a 26-dimensional expression coefficient estimate。 I have an idea： A person had two pictures, one with no expression, the other with exaggerated expression; Based on the 3D mesh estimated by the picture without expression, the expression coefficient of another picture and the final 3D mesh were directly estimated. In this way, we can better estimate the person's 3D facial mesh in real time. thanks for your code。