Intractable Live Free-Viewpoint Video with Haptic Feedback

Shaohui Jiao,Dehao Zhao,Yuzhong Chen,Shengdong Wang,Cong Xie,Bo Zhu,Ali Israr,Li Zhang
DOI: https://doi.org/10.1145/3641521.3664412
2024-01-01
Abstract:We present a novel interactable free-viewpoint video (FVV), which generates photo-realistic and editable volumetric content with high degree of freedom. On the one hand, an enhanced visual hull guided neural representation with higher performance is proposed, an-easy-to-use sparse multi-camera capture system is used, as well as a high performance VH-NeRF is proposed for fast generation of photo-realistic FVV results for live streaming. On the other hand, for monocular panorama video input, high-precision depth and coarse dynamic mesh are calculated, which are used for novel view synthesis of right perspective views with motion parallax on the fly. For better interaction with high realism, dual-channel stereohaptics is implemented and attached on VR headsets to obtain haptic feedback. Last, our FVV solution can do effective compression and transmission on both multi-perspective videos and panorama video with depth data, as well as real-time rendering on consumer-grade hardware. To the best of our knowledge, our work is the first interactable FVV solution with high visual quality and real-time haptic feedback. It is established that the user can do intuitive manipulation for immersive experiences in VR/AR applications.
What problem does this paper attempt to address?