Facial Photogrammetry Pipeline
- État: Closed
- Prix: $1100
- Propositions reçues: 6
- Gagnant: solbergluc
Résumé de concours
render engine for faces on a webapp
Replicate what is shown in the [login to view URL] provided in attachments of this contest.
Build an MVP using Python that meets the requirements:
- detect a face from an image and identify the variant facial features
- filter & detect features of the face, then store those features in a text file
- store new expressions of the face in a diff text file
- convert the 2D data into 3D features using digital photogrammetry
- reconstruct the features and expressions of the still face in 3D with **dense reconstruct** algorithms
- use **spatio-temporal smoothening** algorithms to reconstruct facial expressions.
- present the 3D face to the user, it should mimic the users expressions
Do use any open source libraries that you wish.
More in depth of what the final pipeline shall look like is provided in the attachments. For the sake of this contest you only need to show proof of concept that meets all of the above criteria.