Hand and Face Tracking for a Vision Based Interface
Authors: Vaughan, N.
Journal: Brunel University Department of Information Systems and Computing
Abstract:We aim in this project to tackle the difficulties surrounding the topic area of vision-based image interpretation. We generalise that the most useful interesting objects to detect are people. Considering that the most bold and visually apparent features of people are the hands, face and skin, we develop methods for detecting these in images and live video streams. The detectors are then used by higher level functions to analyse their movements, rotation and feed this back into the system. This movement is further demonstrated as an effective way to control applications without using keyboard and mouse.
Source: Manual