Confident Splatting: Confidence-Based Compression of 3D Gaussian Splatting via Learnable Beta Distributions

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Razlighi, AmirHossein Naghi, Golezani, Elaheh Badali, Kasaei, Shohreh
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909665182351360
author Razlighi, AmirHossein Naghi
Golezani, Elaheh Badali
Kasaei, Shohreh
author_facet Razlighi, AmirHossein Naghi
Golezani, Elaheh Badali
Kasaei, Shohreh
contents 3D Gaussian Splatting enables high-quality real-time rendering but often produces millions of splats, resulting in excessive storage and computational overhead. We propose a novel lossy compression method based on learnable confidence scores modeled as Beta distributions. Each splat's confidence is optimized through reconstruction-aware losses, enabling pruning of low-confidence splats while preserving visual fidelity. The proposed approach is architecture-agnostic and can be applied to any Gaussian Splatting variant. In addition, the average confidence values serve as a new metric to assess the quality of the scene. Extensive experiments demonstrate favorable trade-offs between compression and fidelity compared to prior work. Our code and data are publicly available at https://github.com/amirhossein-razlighi/Confident-Splatting
format Preprint
id arxiv_https___arxiv_org_abs_2506_22973
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Confident Splatting: Confidence-Based Compression of 3D Gaussian Splatting via Learnable Beta Distributions
Razlighi, AmirHossein Naghi
Golezani, Elaheh Badali
Kasaei, Shohreh
Graphics
Computer Vision and Pattern Recognition
3D Gaussian Splatting enables high-quality real-time rendering but often produces millions of splats, resulting in excessive storage and computational overhead. We propose a novel lossy compression method based on learnable confidence scores modeled as Beta distributions. Each splat's confidence is optimized through reconstruction-aware losses, enabling pruning of low-confidence splats while preserving visual fidelity. The proposed approach is architecture-agnostic and can be applied to any Gaussian Splatting variant. In addition, the average confidence values serve as a new metric to assess the quality of the scene. Extensive experiments demonstrate favorable trade-offs between compression and fidelity compared to prior work. Our code and data are publicly available at https://github.com/amirhossein-razlighi/Confident-Splatting
title Confident Splatting: Confidence-Based Compression of 3D Gaussian Splatting via Learnable Beta Distributions
topic Graphics
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2506.22973