A Two Stage Data Compression and Decompression Technique for Point Cloud Data

• Published in: 2018 World Symposium on Digital Intelligence for Systems and Machines (DISA) pp. 297 - 302
• Main Authors: Chakraborty, Subhas, Bandyopadhyay, Avinandan, Yechangunja, Ramesh
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.08.2018
• Subjects: Bandwidth; Cameras; Data compression; Real-time systems; Streaming media; Three-dimensional displays
• DOI: 10.1109/DISA.2018.8490525

The domain of Augmented Reality (AR) and Mixed Reality (MR) that involves capturing and transmitting real time point cloud data, are areas of research for over a decade. However, there are lot of challenges in transmitting huge volume of data over the network because of network bandwidth constraint. This paper proposes a compression strategy for 3D positional data (derived from depth frame) to reduce large volume of captured point cloud data before sending over network to enhance data transmission efficiency, and a decompression technique to be applied at the receiving end to recover the original data. The proposed strategy is well suited for AR/MR interactive applications such as teleconferencing, tele-health monitoring, remote machine monitoring etc. where human subject(s) are the primary data for transmission. In the proposed two stage strategy (hybrid approach), the 3D positional data is first optimized in an in-memory buffer (Buffer Differential technique) without loss of quality. In the second stage, a standard compression algorithm is applied to the output of the first stage for further improving the compression ratio. For this work, we have chosen the Zstandard compression algorithm. The original data is then recovered by applying decompression techniques in the reverse order. A significant improvement has been achieved for this particular class of applications.