It seems that in retinaface (https://github.com/serengil/retinaface), the landmarks "nose", "mouth_right", and "mouth_left" are accessible when'd detecting faces, but I could not see them in the results with deepface, even when using retinaface as backend. Perhaps it would be possible to make those accessible if this backend is used?
Description
It seems that in retinaface (https://github.com/serengil/retinaface), the landmarks "nose", "mouth_right", and "mouth_left" are accessible when'd detecting faces, but I could not see them in the results with deepface, even when using retinaface as backend. Perhaps it would be possible to make those accessible if this backend is used?
Additional Info
No response