Wearable ‘black box' sees you through day

Business Weekly
March 12, 2004

A team from Microsoft Research in Cambridge is developing a personal ‘black box', which digitally and automatically records all the key events in the wearer's day.

Developed by the hardware systems group at MSR Cambridge, led by Lyndsay Williams, SenseCam is a badge-sized wearable camera that captures up to 2000 images per day in a 128Mbyte FLASH memory.

While the concept is still at prototype stage, the team are investigating a host of different applications. Early work centres around the development of both a memory aid and a visual diary. The system was recently showcased at TechFest at Microsoft's Redmond HQ – a get-together of MSR's research teams from around the world.

Sensor data such as movement, light level and temperature is recorded every second and is used intuitively to ensure that the camera captures key events in a day.

Sensors trigger a new recording. For example, each time the person walks into a new room, this light change is detected and the image is captured with an ultra wide angle or fish-eye lens. Other triggers include time, sudden movement or a person nearby.

A hand gesture can also manually capture an image. In order to make the camera wearable – it is worn around the neck – an accelerometer is used for image stabilisation to reduce blurred images caused by camera motion.

The SenseCam can also capture handwritten notes and sketches and transfer the image to PC.

Lyndsay Williams said: “This is similar to an aircraft black box accident recorder, but miniaturised for the human body. It could help with memory recall, reminding the absent-minded wearer, for example where they left their spectacles or keys or reminding them who they met last week, by doing a ‘rewind' of the days events.

“If a person has an accident, the events and images leading up to this will be recorded, and these could be useful to medical staff. It could also be used for automatic diary generation.”

Williams' group is collaborating with another research team in the Bay Area of California, which is developing software that stores, organises and searches data from myriad different sources, including newspaper and magazine articles, home movies, photos, web pages, radio and television – called at this stage, MyLifeBits.

The platform is used to digest and display SenseCam output and integrate it with other data.

The team intends to integrate other technology such as face recognition, so that the wearer can be reminded where they have seen someone before, for example.