VEN-3DVE: vision based egocentric navigation for 3D virtual environments

Navigation is moving inside a virtual environment to reach to a desired object or location. This paper presents a new egocentric navigation technique for exploring a virtual scene by tracking the 3DOF moves of index finger along x, y and z axis. With image processing at backend, the finger’s position in 2D image frame is traced dynamically to travel inside the designed frontend 3D scene. Mimicking a camera in hand, the system can also support Panning. Dynamic mapping is performed to interactively follow the trajectory of index finger. For either of the interactions, movement of index finger controls orientation and position of virtual camera. Linking the libraries of OpenCV and OpenGL, the technique is implemented in a visual studio project navigation by index tracing (NBIT). With NBIT, parallel processing is performed for backend index tracking and front-end rendering of virtual scene. The system was evaluated by fourteen participants for a total of 336 trails whereas applicability of the approach in virtual reality based computer aided design was assessed by five professional designers. Un-biased statistical evaluation of the system reveals that the approach suits well for interactive virtual and augmented reality applications.