Abstract

Context awareness is an important functionality for wearable computers. In particular, the computer should know where the person is in the environment. This paper proposes an image sequence matching technique for the recognition of locations and previously visited places. As in single word recognition in speech recognition, a dynamic programming algorithm is proposed for the calculation of the similarity of different locations. The system runs on a stand alone wearable computer such as a Libretto PC. Using a training sequence a dictionary of locations is created automatically. These locations are then be recognized by the system in realtime using a hatmounted camera.

Citations