# Segmentation-Driven Initialization for Sparse-view 3D Gaussian Splatting

arXiv: https://arxiv.org/abs/2509.11853 · HTML: https://mlsanity.com/paper/2509.11853/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 26 of its 39 references that are in the map.
- First posted on arXiv 15 Sep 2025.

## Builds on (most cited first, 26 of 26)

- [Segment Anything](https://mlsanity.com/paper/2304.02643/) (2023)
- [3D Gaussian Splatting for Real-Time Radiance Field Rendering](https://mlsanity.com/paper/2308.04079/) (2023)
- [Instant Neural Graphics Primitives with a Multiresolution Hash Encoding](https://mlsanity.com/paper/2201.05989/) (2022)
- [DUSt3R: Geometric 3D Vision Made Easy](https://mlsanity.com/paper/2312.14132/) (2023)
- [SuperGlue: Learning Feature Matching with Graph Neural Networks](https://mlsanity.com/paper/1911.11763/) (2019)
- [Grounding Image Matching in 3D with MASt3R](https://mlsanity.com/paper/2406.09756/) (2024)
- [RegNeRF: Regularizing Neural Radiance Fields for View Synthesis from Sparse Inputs](https://mlsanity.com/paper/2112.00724/) (2021)
- [Gaussian Grouping: Segment and Edit Anything in 3D Scenes](https://mlsanity.com/paper/2312.00732/) (2023)
- [FSGS: Real-Time Few-shot View Synthesis using Gaussian Splatting](https://mlsanity.com/paper/2312.00451/) (2023)
- [DNGaussian: Optimizing Sparse-View 3D Gaussian Radiance Fields with Global-Local Depth Normalization](https://mlsanity.com/paper/2403.06912/) (2024)
- [MVImgNet: A Large-scale Dataset of Multi-view Images](https://mlsanity.com/paper/2303.06042/) (2023)
- [SPARF: Neural Radiance Fields from Sparse and Noisy Poses](https://mlsanity.com/paper/2211.11738/) (2022)
- [COLMAP-Free 3D Gaussian Splatting](https://mlsanity.com/paper/2312.07504/) (2023)
- [Stereo Radiance Fields (SRF): Learning View Synthesis for Sparse Views of Novel Scenes](https://mlsanity.com/paper/2104.06935/) (2021)
- [VR-GS: A Physical Dynamics-Aware Interactive Gaussian Splatting System in Virtual Reality](https://mlsanity.com/paper/2401.16663/) (2024)
- [Geometry-Aware Learning of Maps for Camera Localization](https://mlsanity.com/paper/1712.03342/) (2017)
- [SparseGS: Sparse View Synthesis using 3D Gaussian Splatting](https://mlsanity.com/paper/2312.00206/) (2023)
- [CoR-GS: Sparse-View 3D Gaussian Splatting via Co-Regularization](https://mlsanity.com/paper/2405.12110/) (2024)
- [Progressively Optimized Local Radiance Fields for Robust View Synthesis](https://mlsanity.com/paper/2303.13791/) (2023)
- [Superpixel Segmentation with Fully Convolutional Networks](https://mlsanity.com/paper/2003.12929/) (2020)
- [InstantSplat: Sparse-view Gaussian Splatting in Seconds](https://mlsanity.com/paper/2403.20309/) (2024)
- [3D Gaussian Splatting in Robotics: A Survey](https://mlsanity.com/paper/2410.12262/) (2024)
- [SAM3D: Zero-Shot 3D Object Detection via Segment Anything Model](https://mlsanity.com/paper/2306.02245/) (2023)
- [EndoGS: Deformable Endoscopic Tissues Reconstruction with Gaussian Splatting](https://mlsanity.com/paper/2401.11535/) (2024)
- [TiAVox: Time-aware Attenuation Voxels for Sparse-view 4D DSA Reconstruction](https://mlsanity.com/paper/2309.02318/) (2023)
- [Adaptive Segmentation-Based Initialization for Steered Mixture of Experts Image Regression](https://mlsanity.com/paper/2409.10101/) (2024)

## Abstract (arXiv)

Sparse-view synthesis remains a challenging problem due to the difficulty of recovering accurate geometry and appearance from limited observations. While recent advances in 3D Gaussian Splatting (3DGS) have enabled real-time rendering with competitive quality, existing pipelines often rely on Structure-from-Motion (SfM) for camera pose estimation, an approach that struggles in genuinely sparse-view settings. Moreover, several SfM-free methods replace SfM with multi-view stereo (MVS) models, but generate massive numbers of 3D Gaussians by back-projecting every pixel into 3D space, leading to high memory costs. We propose Segmentation-Driven Initialization for Gaussian Splatting (SDI-GS), a method that mitigates inefficiency by leveraging region-based segmentation to identify and retain only structurally significant regions. This enables selective downsampling of the dense point cloud, preserving scene fidelity while substantially reducing Gaussian count. Experiments across diverse benchmarks show that SDI-GS reduces Gaussian count by up to 50% and achieves comparable or superior rendering quality in PSNR and SSIM, with only marginal degradation in LPIPS. It further enables faster training and lower memory footprint, advancing the practicality of 3DGS for constrained-view scenarios.