Can Compact 3D Descriptors Enhance Semantic Face Compression in the Metaverse?

Original title: Semantic Face Compression for Metaverse: A Compact 3D Descriptor Based Approach

Authors: Binzhe Li, Bolin Chen, Zhao Wang, Shiqi Wang, Yan Ye

The article introduces a novel approach for virtual avatar communication in the metaverse—semantic face compression using compact 3D facial descriptors. Emphasizing conveying semantic information in avatar faces, this method stands out for its flexibility, efficiency, and semantic richness. By enabling communication of descriptors for AI comprehension, it opens doors to diverse metaverse applications sans human involvement. Performance comparisons against Versatile Video Coding reveal remarkable rate-accuracy advancements. This paradigm shift not only enhances digital human communication through machine analysis but also promises to revolutionize interaction and communication within the metaverse. Its potential spans wide, envisioning a future where avatars convey nuanced meanings and drive diverse applications, shaping a new era of metaverse interaction.

Original article: https://arxiv.org/abs/2311.12817