Stanford University

Temporal RVL: A Depth Stream Compression Method

Jun, H., & Bailenson, J.N. (2020). Temporal RVL: A Depth Stream Compression Method. Proceedings of IEEE Virtual Reality (VR)

View PDF Download Software

Abstract

The advent of depth cameras has led to new opportunities, but at the same time has led to new challenges in the form of larger network bandwidth. To address this problem, we propose a lossy compression method Temporal RVL, which results in better compression with little loss of depth information. Temporal RVL adds a preprocessing step to RVL and effectively utilizes the similarities across frames, while maintaining important depth information such as edges. For the default settings, Temporal RVL achieves a compression ratio of 20.1 (4.2 times higher than RVL) while at the same time facilitating faster decompression.

Comments are closed.