Hi YadirF, I'd like to use your code for the 3D reconstruction of provided 2D single image.
But your code starts from an already provided 3D model which holds vertices and expression information in 3D coordinate format.
Do you think would it be possible for me to use your PRnet for pre-fitting of single image into 3D then feed it into your face3d to find the PCA coefficients?
It looks like your PRnet presents slightly less vertices(40K) compared to face3d(50K) and I hope there semantical point representation has some consistency.
I presumably guess if I can load the 300LP face model into the face3d, it might work cuz you denoted that you are using this model for your PRnet. How do you think?
Hi YadirF, I'd like to use your code for the 3D reconstruction of provided 2D single image. But your code starts from an already provided 3D model which holds vertices and expression information in 3D coordinate format.
Do you think would it be possible for me to use your PRnet for pre-fitting of single image into 3D then feed it into your face3d to find the PCA coefficients?
It looks like your PRnet presents slightly less vertices(40K) compared to face3d(50K) and I hope there semantical point representation has some consistency.
I presumably guess if I can load the 300LP face model into the face3d, it might work cuz you denoted that you are using this model for your PRnet. How do you think?