A Systematic Survey Into Compression Algorithms for Three-Dimensional Content

Abstract

This systematic review investigates compression algorithms for three-dimensional content, focusing on recent advancements. It categorizes the methodologies into traditional, learning-based, and semantic approaches. The review includes 52 studies selected based on criteria including publication date, peer review status, and relevance to the field. The analysis highlights the significant contributions of learning-based and semantic techniques in advancing 3D content compression. Notably, some reviewed learning-based methods demonstrated over 95% improvement in rate optimization compared to standard point cloud compression methods. Despite the comprehensive coverage, the review acknowledges certain limitations due to potential biases in study selection and the inherent heterogeneity of the included research. The findings underscore the importance of continued exploration in learning-based and semantic compression for enhancing the efficiency and applicability of 3D content technologies.