Toshiba shows advanced image detection prototypes

Toshiba is using this week's IFA electronics show in Berlin to show off some of its latest image detection technology currently being worked on at its laboratories in Cambridge in the UK.

One example is a gesture control system that allows for interaction with a TVs interface through hand movements.

It watches for a person to come into its field of view and, once recognized, looks for their hand. Once the hand is identified, the user just makes a fist and can then wave it around in the air as if controlling a mouse. A cursor on the TV interface moves across the screen mirroring the user's fist movements.

The gesture control prototype is being shown a few steps away from a display showcasing Toshiba's new Qosmio laptops that are the first products to include some of the same technology. The laptops feature an earlier version of the system that allows for limited interaction through hand gestures but works in much the same way. Users make a fist to move a cursor on the screen and then raise their thumb to perform a mouse click.

"The major difference with the Qosmio is that it's just smoother and more advanced tracking technology," said Kate Knill, manager of interaction technology at the laboratory, speaking of the research prototype on show. It is also much better at picking out a single user from a crowd and keeping locked on them rather than getting confused by the hands of other people in its picture, she said.

Having developed the prototype to this stage, Toshiba is working on ensuring it will function in real-world situations and sees one possible use for the technology as a secondary interface to a TV in addition to the conventional remote control.

A second video-based system is a pattern recognition system that has a video camera mounted above a TV screen watching for a card -- in this case a German or British flag -- as a cue to change the language on a video that's playing on the TV.

Toshiba sees several possible applications for the technology in the future, including, for example, a children's learning game where kids have to find the correct card in response to a question or instruction on-screen.

The same system could be used to recognize TV viewers and present them with personalized information or switch to their favorite channel when they walk in the room, said Knill.