Smiles less noise

Video
Apr 1–30, 2008
Sib.tv, Shibuya

Production of video work as part of the Taico Club Art Project project.
In this video, the facial expression is obtained using a myoelectric potential sensor.
Using a pattern recognition model called Support Vector Machine, facial expressions were classified into three patterns: expressionless, smiling, and angry.
We have developed a video effect that reduces noise when approaching a smile and increases noise when approaching anger.

It was supposed to be aired once an hour at the largest intersection in Japan, with 800,000 passengers per day, but my video was reported to the police (noise?) And the screening was canceled and only once happened.

Concept, visual, sound and programming: Daito Manabe
Sensor development: Masaki Teruoka