Transcoding a 3D Gaussian Splatting Model from a Plenoptic Point Cloud or Mesh without the Original Multi-view Images

TL;DR

This paper introduces an end-to-end pipeline for converting 3D point clouds or meshes into 3D Gaussian splatting models without needing original multi-view images, improving efficiency and quality.

Contribution

The authors present a novel transcoding method with custom initialization that enhances convergence speed and surface quality in 3D Gaussian splatting models.

Findings

Produces high-quality 3D Gaussian splatting models with fewer splats than original points.

Custom initialization significantly speeds up convergence and improves surface clarity.

Effective on standard plenoptic point cloud datasets.

Abstract

In this paper, we propose an end-to-end transcoding pipeline, to create 3D Gaussian splatting (3DGS) models from existing 3D plenoptic point cloud or mesh models, when the original multi-view images of the captured 3D object or scene are not available. We also propose a custom initialisation to guide the 3DGS model learning, with constraints to ensure that the final 3DGS model aligns closely with the input point cloud or mesh surface. Tests on a high-quality, standard plenoptic point cloud dataset show that our pipeline produces 3DGS models of high visual quality, with many fewer splats than points in the original dense point clouds. Additionally, our custom initialisation leads to much faster convergence and cleaner surface representation than when starting from the default SfM-based initialisation that is typically used for 3DGS model learning.