Interactive Control through Hand Gestures

Abstract

In recent years, we have seen dramatic progress in motion capture techniques for digital character animation, yet these techniques are expensive due to the need of special hardware. Keeping in view the trend of making applications interactive and more natural and intuitive, as well as keeping the cost low, a vision-based system is developed to allow interactive control of applications, using hand gestures. The system is used for two types of applications: skeleton animation and image editing. The procedure steps mainly consist background subtraction, fingertips detection, tracking, and mapping of finger lengths to control parameters. For skeleton animation application, finger lengths are mapped to joint angles. For image editing application, finger lengths are mapped to parameters such as rotation angle of image or parameters for image cropping. In addition to the idea of using hand gestures for skeleton animation and image editing, we have introduced a novel method for detecting fingertips, in which the hand contour is traversed and a turning point is found using trailing buffer.