Occlusion for 3D Object Manipulation with Hands in Augmented Reality

By dynamically enhancing the reality with a wide range of applications, augmented reality has been drawing growing attention during recent years. Due to the need to interact with virtual objects, hand-object interaction has become a critical element in these applications. By enhancing existing objects with virtual visual appearance, tactile feedback can be provided to the user as well. However, to deal with problems of augmenting real objects with virtual data, a correct visual representation is needed. This becomes even more critical when the interaction between virtual and real objects happens. Incorrect occlusion information can be fatal to an immersive augmented reality experience. In this paper, we propose an innovative approach to handle the occlusion of augmented 3D object manipulation with hands in real-time by exploiting the nature of hand pose combined with tracking-based and model-based methods for estimating occlusion, to achieve a complete augmented reality experience without the necessities of heavy computations, complex manual segmentation processes or wearing special gloves.