Abstract

This presentation will be of interest to 1) physical activity researchers interested in better capturing participant physical activity and sedentary behaviour, and 2) computer science researchers interested in wearable sensors and image processing challenges in a new application area. In this lecture, Dr Doherty will discuss use of the Microsoft Research SenseCam, a wearable camera sensor which automatically captures up to 5,000 first-person point-of-view images. The images strongly help identify the type of activity a participant is involved in, and can also help determine the environment and situation surrounding the given activity. Because up to 35,000 images can be captured from a participant each week, there exists a substantial information management challenge in storing, annotating, and retrieving image content. This talk will discuss current state-of-art computational approaches applied to SenseCam images, and their application in the field of physical activity.