Face Recognition Games
Banksy in us
Banksy in us is a project that prints out a holiday theme postcard with a photo image of a user.
By face recognition, users are encouraged to make funny faces that match popular gif images.
To compare the facial expression, extracted facial points using FaceOSC, and compared the ratio of coordinates.
When the app takes a photo image, users can make a postcard image that looks like Banksy graffitis.
The final selected postcard image is also printed out as a souvenir, and it can be uploaded to Instagram account if they agree with.
Algorithm
Sample clips are pre-analyzed with the coordinates of key facial points
[Orientation coordinate(x,y,z), Eyebrow heights(left, right), Eye heights(left, right), nose coordinate, mouth width and height].
These data are stored in the app, and compared to the user face coordinated to compare the facial expression similarity.
The app does not simply compare the coordinates; it compares the ratio of components of the face and the orientation point. This algorithm is devised to deal with different facial structures and the distance from the camera.
When the score based on the similarity approaches the threshold, the camera takes a photo of users.
UX flow
System
FacsOSC: Face Recognition Processing: Whole Structure
Webcam: Face image Input
Printer: Postcard Output
Role
Programming, System ArchitectureCredit
Namsoo Kim: UX design, DocumentationFeatured at ITP Winter Show 2017
Winter 2017
Winter 2017
iCeleb
System
FacsOSC: Face Recognition Processing: Whole Structure
Webcam: Face image Input
Role
System Architecture, ProgrammingCredit
Tuangkamol Thongborisute: Concept, UX Design, DocumentationFeatured at UCLA Design|Media Arts Solo Show by Tuang Thongborisute
Los Angeles, Spring 2017
Los Angeles, Spring 2017