A virtual character recognition system based on finger tracking is proposed.By capturing the finger movement with a digital camera and tracking the finger-tip position automatically

the 2-D character image (virtual character trajectory) "written" by finger-tip is reconstructed and recognized.A modified flat functional link neural network is used to predict and track the moving position of the finger-tip.A mathematical morphology and template matching based finger-tip detection approach is proposed

and an algorithm to reconstruct and recognize the virtual character is given.Experiments show that the proposed system works very well with high finger tracking and detection accuracy

and achieves more than 95% accuracy rate for virtual digital and English letters recognition.