V-PCC Performance Evaluation of the First MPEG Point Codec

Abstract

Representation of 3D scenes has become increasingly important in several industries using virtual and augmented reality technologies. The point cloud format is well suited for such representations. Indeed, point clouds can be created with a simple capture process and modest processing, enabling a realtime, end-to-end point cloud distribution chain. However, point cloud compression is required to obtain data rates and file sizes that could be economically viable for the industry. Standardization is required to ensure interoperability. In 2021, the International Organization for Standardization (ISO) Moving Picture Experts Group (MPEG) plans to publish a standard for its first point cloud codec, MPEG-I Part 5: Visual Volumetric Video-based Coding (V3C) and Video-based Point Cloud Compression (V-PCC). This standard enables a world of new services and applications, including cultural heritage, telepresence, and new forms of entertainment. In this paper, we review the principal use cases targeted by the V-PCC standard. We present the architecture of the V-PCC codec and describe its main tools, by giving insights into complexity at the encoder and decoder levels and explaining profiles and conformance points in V-PCC. We then present the methodology established, as a collaboration between industry and academics, for the evaluation of the V-PCC codec performance and the methodology’s origins. This methodology was applied to the MPEG point cloud compression test model software (named TMC2) to consistently evaluate technologies proposed during the standardization process. Finally, we compare the performance of the main V-PCC tools available for lossy and lossless compression. Finally, the conclusion provides elements in favor of a near-term deployment of V-PCC in the media industry \n.