Enregistré dans:
Détails bibliographiques
Auteurs principaux: Xin, Yuelin, Liu, Yuheng, Xie, Xiaohui, Li, Xinke
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:https://arxiv.org/abs/2509.22917
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866918316237389824
author Xin, Yuelin
Liu, Yuheng
Xie, Xiaohui
Li, Xinke
author_facet Xin, Yuelin
Liu, Yuheng
Xie, Xiaohui
Li, Xinke
contents 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. Implementation available at https://github.com/cilix-ai/gs-embedding
format Preprint
id arxiv_https___arxiv_org_abs_2509_22917
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Learning Unified Representation of 3D Gaussian Splatting
Xin, Yuelin
Liu, Yuheng
Xie, Xiaohui
Li, Xinke
Computer Vision and Pattern Recognition
I.4.10
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. Implementation available at https://github.com/cilix-ai/gs-embedding
title Learning Unified Representation of 3D Gaussian Splatting
topic Computer Vision and Pattern Recognition
I.4.10
url https://arxiv.org/abs/2509.22917