A robust hand detection and tracking algorithm with application to natural user interface

This paper describes a system for hand detection and tracking which is used to implement a natural user interface which allows controlling the computer with hand movements and gestures. Different image processing and computer vision techniques were used in the realization of this system. The main components of the system are a skin detector based on histograms whose results have been improved with background subtraction and a clustering algorithm, an open hand detector and a subsystem for object tracking with a flock of features algorithm. The system has proved to be very robust to different background and lightning conditions and can be used in a wide array of multimedia and desktop applications.