Closed 12345data closed 4 years ago
The filename_lmks.npy is just an array with 51 facial landmarks (i.e. a 51 x 3 dimensional matrix). To obtain such a file, run for instance the FAN landmark predictor on the image that you want to fit and save the landmarks in a Numpy file, excluding the first 17 landmarks (i.e. the boundary keypoints). Does this answer your question?
Yes. Thanks for the clarification and fast reply. Tried the same and got the landmark file.
On Thu, Sep 17, 2020 at 12:24 AM TimoBolkart notifications@github.com wrote:
The filename_lmks.npy is just an array with 51 facial landmarks (i.e. a 51 x 3 dimensional matrix). To obtain such a file, run for instance the FAN landmark predictor https://github.com/1adrianb/face-alignment on the image that you want to fit and save the landmarks in a Numpy file, excluding the first 17 landmarks (i.e. the boundary keypoints). Does this answer your question?
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/TimoBolkart/TF_FLAME/issues/31#issuecomment-693596099, or unsubscribe https://github.com/notifications/unsubscribe-auth/AELCCCIFSONADQNYLAYWZVTSGECXXANCNFSM4RL6PDAQ .
How filename_lmks.npy files are created. I was trying to create a textured obj file for voca. When I use the files present in TF_FLAME i was enable to create the textured model but failed while trying with npy files which was created from the Ringnet. When I checked the content of lmks.npy file and the npy files which was created in Ringnet, they were totally different. How can I create these npy files. What all things I need to do to make the textured model to animate for talking my sentences?