View-Dependent Virtual Reality Content from RGB-D Images
[Technical Report, Research Paper]
UNIVERSITY OF SOUTHERN CALIFORNIA LOS ANGELES
Pagination or Media Count:
High-fidelity virtual content is essential for the creation of compelling and effective virtual reality VR experiences. However, creating photorealistic content is not easy, and handcrafting detailed 3D models can be time and labor intensive. Structured camera arrays, such as light-stages, can scan and reconstruct high-fidelity virtual models, but the expense makes this technology impractical for most users. In this paper, we present a complete end-to-end pipeline for the capture, processing, and rendering of view-dependent 3D models in virtual reality from a single consumer-grade depth camera. The geometry model and the camera trajectories are automatically reconstructed and optimized from a RGB-D image sequence captured offline. Based on the head-mounted display HMD position, the three closest images are selected for real-time rendering and fused together to smooth the transition between viewpoints. The specular reflections and light-burst effects can also be preserved and reproduced. We confirmed that our method does not require technical background knowledge by testing our system with data captured by non-expert operators.
- Computer Programming and Software