Matryoshka Gaussian Splatting

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Guo, Zhilin, Zhang, Boqiao, Aktas, Hakan, Fogarty, Kyle, Hu, Jeffrey, Aslan, Nursena Koprucu, Li, Wenzhao, Baykal, Canberk, Miao, Albert, Bengtson, Josef, Zhou, Chenliang, Xia, Weihao, Vasconcelos, Cristina Nader, Oztireli, Cengiz
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910061169737728
author Guo, Zhilin
Zhang, Boqiao
Aktas, Hakan
Fogarty, Kyle
Hu, Jeffrey
Aslan, Nursena Koprucu
Li, Wenzhao
Baykal, Canberk
Miao, Albert
Bengtson, Josef
Zhou, Chenliang
Xia, Weihao
Vasconcelos, Cristina Nader
Oztireli, Cengiz
author_facet Guo, Zhilin
Zhang, Boqiao
Aktas, Hakan
Fogarty, Kyle
Hu, Jeffrey
Aslan, Nursena Koprucu
Li, Wenzhao
Baykal, Canberk
Miao, Albert
Bengtson, Josef
Zhou, Chenliang
Xia, Weihao
Vasconcelos, Cristina Nader
Oztireli, Cengiz
contents The ability to render scenes at adjustable fidelity from a single model, known as level of detail (LoD), is crucial for practical deployment of 3D Gaussian Splatting (3DGS). Existing discrete LoD methods expose only a limited set of operating points, while concurrent continuous LoD approaches enable smoother scaling but often suffer noticeable quality degradation at full capacity, making LoD a costly design decision. We introduce Matryoshka Gaussian Splatting (MGS), a training framework that enables continuous LoD for standard 3DGS pipelines without sacrificing full-capacity rendering quality. MGS learns a single ordered set of Gaussians such that rendering any prefix, the first k splats, produces a coherent reconstruction whose fidelity improves smoothly with increasing budget. Our key idea is stochastic budget training: each iteration samples a random splat budget and optimises both the corresponding prefix and the full set. This strategy requires only two forward passes and introduces no architectural modifications. Experiments across four benchmarks and six baselines show that MGS matches the full-capacity performance of its backbone while enabling a continuous speed-quality trade-off from a single model. Extensive ablations on ordering strategies, training objectives, and model capacity further validate the designs.
format Preprint
id arxiv_https___arxiv_org_abs_2603_19234
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Matryoshka Gaussian Splatting
Guo, Zhilin
Zhang, Boqiao
Aktas, Hakan
Fogarty, Kyle
Hu, Jeffrey
Aslan, Nursena Koprucu
Li, Wenzhao
Baykal, Canberk
Miao, Albert
Bengtson, Josef
Zhou, Chenliang
Xia, Weihao
Vasconcelos, Cristina Nader
Oztireli, Cengiz
Computer Vision and Pattern Recognition
Graphics
The ability to render scenes at adjustable fidelity from a single model, known as level of detail (LoD), is crucial for practical deployment of 3D Gaussian Splatting (3DGS). Existing discrete LoD methods expose only a limited set of operating points, while concurrent continuous LoD approaches enable smoother scaling but often suffer noticeable quality degradation at full capacity, making LoD a costly design decision. We introduce Matryoshka Gaussian Splatting (MGS), a training framework that enables continuous LoD for standard 3DGS pipelines without sacrificing full-capacity rendering quality. MGS learns a single ordered set of Gaussians such that rendering any prefix, the first k splats, produces a coherent reconstruction whose fidelity improves smoothly with increasing budget. Our key idea is stochastic budget training: each iteration samples a random splat budget and optimises both the corresponding prefix and the full set. This strategy requires only two forward passes and introduces no architectural modifications. Experiments across four benchmarks and six baselines show that MGS matches the full-capacity performance of its backbone while enabling a continuous speed-quality trade-off from a single model. Extensive ablations on ordering strategies, training objectives, and model capacity further validate the designs.
title Matryoshka Gaussian Splatting
topic Computer Vision and Pattern Recognition
Graphics
url https://arxiv.org/abs/2603.19234