Virtual programming with bare-hand-based interaction

Virtual programming is the front end of virtual assembly. It provides the interface for product engineers to plan, evaluate, optimize, and verify the assembly of mechanical products without the need of physical objects. However, the dependency of virtual programming on specialized virtual reality devices, such as data gloves, imposes limits on the convenience of human-computer interaction. Presented in this paper is a new approach of bare-hand-based posture recognition and motion tracking. It localizes and segments fingertips with a combined method of skin detection and contour retrieving, thus producing accurate results. Experiments demonstrate the robustness of the presented approach and its use in the interaction design for virtual programming.