Learning Unified Representation of 3D Gaussian Splatting

Yuelin Xin, Yuheng Liu, Xiaohui Xie, Xinke Li

Keywords: Representation Learning, 3D Gaussian Splatting

TL;DR: Proposed a new representation of 3DGS based on submanifold field that is more suitable for learning.

Abstract: A well-designed vectorized representation is crucial for the learning systems natively based on 3D Gaussian Splatting. While 3DGS enables efficient and explicit 3D reconstruction, its parameter-based representation remains hard to learn as features, especially for neural-network-based models. Directly feeding raw Gaussian parameters into learning frameworks fails to address the non-unique and heterogeneous nature of the Gaussian parameterization, yielding highly data-dependent models. This challenge motivates us to explore a more principled approach to represent 3D Gaussian Splatting in neural networks that preserves the underlying color and geometric structure while enforcing unique mapping and channel homogeneity. In this paper, we propose an embedding representation of 3DGS based on continuous submanifold fields that encapsulate the intrinsic information of Gaussian primitives, thereby benefiting the learning of 3DGS.

Primary Area: unsupervised, self-supervised, semi-supervised, and supervised representation learning

Submission Number: 2196