SurFace
In Short
SurFace is a spatial interactive installation that investigates biometric data extraction, identity reductionism, and algorithmic bias. Using real-time facial recognition and demographic inference models, the installation scans viewers in front of a vertical mirror display and distributes its real-time classifications: gender, age, ethnicity, and emotional valence, across surrounding satellite screens.
Ft: Pili Valdivia
Tech Stack
- Live Ingestion & Optical Tracking: A vertical central screen functions as an algorithmic mirror. An overhead camera captures audience movement.
- Biometric Pipeline (DeepFace + TouchDesigner):
- Video streams feed into DeepFace to classify facial attributes against trained datasets.
- Real-time outputs predict four discrete dimensions: Gender, Age, Ethnicity, and Emotional State.
- Decoupled Multi-Screen Spatialization
- Dynamic Audio Feedback (Ableton Live 12): System-triggered sonic cues respond to tracking locks and classification shifts
Installation Views
