BlinkI uses eye blinks performed to the rhythm of songs for signalling messages or controlling objects. The eyeblinks are detected using a standard webcam. Such a system could be used by someone who is paralyzed to trigger pre-trained messages or control objects in their environment. We also explored using eyeblinks to control wearable computers using a proximity sensor mounted next to a head-up display to detect the eyeblinks. Using shorter, more subtle patterns, the wearable computer user can trigger actions on his computer without interrupting a conversation or bothering bystanders. Surprisingly, eyeblinks can be used as a biometric to verify the identity of the blinker. Imagine a door security system with a video camera where a person who wants to gain entry walks up to the system and blinks their personal pattern (similar to a PIN). It is more secure than typing in a PIN on a number pad because eyeblinking is more difficult for a bystander to see. However, even if an intruder knew the pattern, individual differences in the motion of a person's eyeblinks would alert the security system to the intruder even though he performed the correct pattern. Combining the system with face recognition further improves security. This demo was created by Joseph Park to illustrate the concept while he was a post-doc at Professor Thad Starner's Contextual Computing Group. For more details, please see the paper "Recognizing Song--Based Blink Patterns: Applications for Restricted and ...
No comments:
Post a Comment