Real Time Gesture Learning and Recognition: Towards Automatic Categorization

This research focuses on real-time gesture learning and recognition. Events arrive in a continuous stream without explicitly given boundaries. To obtain temporal accuracy, we need to consider the lag between the detection of an event and any effects we wish to trigger with it. Two methods for real time gesture recognition using a Nintendo Wii controller are presented. The first detects gestures similar to a given template using either a Euclidean distance or a cosine similarity measure. The second method uses novel information theoretic methods to detect and categorize gestures in an unsupervised way. The role of supervision, detection lag and the importance of haptic feedback are discussed.