Hand Segmentation for Augmented Reality System

A hand segmentation method for augmented reality system had been proposed in this paper. Color-based and appearance-based skin detection technologies both have been employed in this method to separate the hand from the background. There are two benefits of our method in comparison with the traditional segmentation method: one is that it does not require a stationary camera or a static background; another is that it is not sensitive to intensity different, and requires no special object model with relative high performance. Thus, our method is applicable to the augmented reality systems or other real-time systems. Due to apply an appearance-based skin detection technology to reduce the misjudged pixels, the quality of the segmentation result is better than traditional color-based skin detection techniques. Finally, some experiments had been carried out to assess the accuracy, performance and robustness of our method. The results show the effectiveness of our method.