Ever wonder who got more face time: Captain Kirk or Spock? We know it wasn’t Chekov…

PittPatt – a spin-off company of Carnegie Mellon – has an answer. The company’s “Face Mining” technology can detect Captain Kirk’s face from the front and slightly tilted. Soon, it will be able to detect faces from other angles. It then records every scene featuring Kirk and let’s a user navigate all three seasons of Star Trek by not only a favorite episode (the Trouble with Tribbles anyone?) but also a favorite character.

The implications of this technology for other uses – such as security monitoring – are huge.

Face Mining works by combining the algorithms used in face detection, face tracking and face recognition to extract all visible face tracks and group images of the same person together.

PittPatt – Pittsburgh Pattern Recognition – is operated by alumni Henry Schneiderman (E’90, CS’00), Michael Nechyba (CS’98) and Michael Sipe (E’99).


Anonymous said...

Cool stuff!!

Sally said...

How would this work for an episode where one character impersonates another? There are lots of Star Trek Voyager episodes where characters are genetically altered or wearing disguises.

Marco said...

My new Canon digicam can detect faces in a scene - would be cool if it could auto-tag names to who's in the photos.

Can this technology compensate for age, say over an actor's 30-year career, or as a family member grows older?