# SfM-Free 3D Gaussian Splatting via Hierarchical Training

arXiv: https://arxiv.org/abs/2412.01553 · HTML: https://mlsanity.com/paper/2412.01553/

## Key facts

- Cited by 12 papers in the map (data through 14 Oct 2026).
- Builds on 31 of its 40 references that are in the map.
- First posted on arXiv 2 Dec 2024.

## Builds on (most cited first, 30 of 31)

- [The Unreasonable Effectiveness of Deep Features as a Perceptual Metric](https://mlsanity.com/paper/1801.03924/) (2018)
- [3D Gaussian Splatting for Real-Time Radiance Field Rendering](https://mlsanity.com/paper/2308.04079/) (2023)
- [Mip-NeRF 360: Unbounded Anti-Aliased Neural Radiance Fields](https://mlsanity.com/paper/2111.12077/) (2021)
- [Mip-NeRF: A Multiscale Representation for Anti-Aliasing Neural Radiance Fields](https://mlsanity.com/paper/2103.13415/) (2021)
- [Vision Transformers for Dense Prediction](https://mlsanity.com/paper/2103.13413/) (2021)
- [Plenoxels: Radiance Fields without Neural Networks](https://mlsanity.com/paper/2112.05131/) (2021)
- [4D Gaussian Splatting for Real-Time Dynamic Scene Rendering](https://mlsanity.com/paper/2310.08528/) (2023)
- [Mip-Splatting: Alias-free 3D Gaussian Splatting](https://mlsanity.com/paper/2311.16493/) (2023)
- [ZoeDepth: Zero-shot Transfer by Combining Relative and Metric Depth](https://mlsanity.com/paper/2302.12288/) (2023)
- [Common Objects in 3D: Large-Scale Learning and Evaluation of Real-life 3D Category Reconstruction](https://mlsanity.com/paper/2109.00512/) (2021)
- [BARF: Bundle-Adjusting Neural Radiance Fields](https://mlsanity.com/paper/2104.06405/) (2021)
- [INeRF: Inverting Neural Radiance Fields for Pose Estimation](https://mlsanity.com/paper/2012.05877/) (2020)
- [Compact 3D Gaussian Representation for Radiance Field](https://mlsanity.com/paper/2311.13681/) (2023)
- [Compressed 3D Gaussian Splatting for Accelerated Novel View Synthesis](https://mlsanity.com/paper/2401.02436/) (2023)
- [NoPe-NeRF: Optimising Neural Radiance Field with No Pose Prior](https://mlsanity.com/paper/2212.07388/) (2022)
- [A Hierarchical 3D Gaussian Representation for Real-Time Rendering of Very Large Datasets](https://mlsanity.com/paper/2406.12080/) (2024)
- [Self-Calibrating Neural Radiance Fields](https://mlsanity.com/paper/2108.13826/) (2021)
- [COLMAP-Free 3D Gaussian Splatting](https://mlsanity.com/paper/2312.07504/) (2023)
- [CityGaussian: Real-time High-quality Large-Scale Scene Rendering with Gaussians](https://mlsanity.com/paper/2404.01133/) (2024)
- [Compact 3D Scene Representation via Self-Organizing Gaussian Grids](https://mlsanity.com/paper/2312.13299/) (2023)
- [Tri-MipRF: Tri-Mip Representation for Efficient Anti-Aliasing Neural Radiance Fields](https://mlsanity.com/paper/2307.11335/) (2023)
- [Multi-Scale 3D Gaussian Splatting for Anti-Aliased Rendering](https://mlsanity.com/paper/2311.17089/) (2023)
- [Revising Densification in Gaussian Splatting](https://mlsanity.com/paper/2404.06109/) (2024)
- [Pixel-GS: Density Control with Pixel-aware Gradient for 3D Gaussian Splatting](https://mlsanity.com/paper/2403.15530/) (2024)
- [IFRNet: Intermediate Feature Refine Network for Efficient Frame Interpolation](https://mlsanity.com/paper/2205.14620/) (2022)
- [Progressively Optimized Local Radiance Fields for Robust View Synthesis](https://mlsanity.com/paper/2303.13791/) (2023)
- [SiNeRF: Sinusoidal Neural Radiance Fields for Joint Pose Estimation and Scene Reconstruction](https://mlsanity.com/paper/2210.04553/) (2022)
- [SWinGS: Sliding Windows for Dynamic 3D Gaussian Splatting](https://mlsanity.com/paper/2312.13308/) (2023)
- [LU-NeRF: Scene and Pose Estimation by Synchronizing Local Unposed NeRFs](https://mlsanity.com/paper/2306.05410/) (2023)
- [A Construct-Optimize Approach to Sparse View Synthesis without Camera Pose](https://mlsanity.com/paper/2405.03659/) (2024)

## Built on by (most cited first, 12 of 12)

- [LongSplat: Robust Unposed 3D Gaussian Splatting for Casual Long Videos](https://mlsanity.com/paper/2508.14041/) (2025)
- [VGGT-X: When VGGT Meets Dense Novel View Synthesis](https://mlsanity.com/paper/2509.25191/) (2025)
- [TrackGS: Optimizing COLMAP-Free 3D Gaussian Splatting with Global Track Constraints](https://mlsanity.com/paper/2502.19800/) (2025)
- [On-the-fly Large-scale 3D Reconstruction from Multi-Camera Rigs](https://mlsanity.com/paper/2512.08498/) (2025)
- [LighthouseGS: Indoor Structure-aware 3D Gaussian Splatting for Panorama-Style Mobile Captures](https://mlsanity.com/paper/2507.06109/) (2025)
- [CompSplat: Compression-aware 3D Gaussian Splatting for Real-world Video](https://mlsanity.com/paper/2602.09816/) (2026)
- [Enhancing Novel View Synthesis from extremely sparse views with SfM-free 3D Gaussian Splatting Framework](https://mlsanity.com/paper/2508.15457/) (2025)
- [Rad-GS: Radar-Vision Integration for 3D Gaussian Splatting SLAM in Outdoor Environments](https://mlsanity.com/paper/2511.16091/) (2025)
- [GloSplat: Joint Pose-Appearance Optimization for Faster and More Accurate 3D Reconstruction](https://mlsanity.com/paper/2603.04847/) (2026)
- [Pose-Free Omnidirectional Gaussian Splatting for 360-Degree Videos with Consistent Depth Priors](https://mlsanity.com/paper/2603.23324/) (2026)
- [TurboGS: Accelerating 3D Gaussian Splatting via Error-Guided Sparse Pixel Sampling and Optimization](https://mlsanity.com/paper/2606.15924/) (2026)
- [TruncGradGS: Improved 3D Gaussian Splatting via Truncated Gradient Updates](https://mlsanity.com/paper/2609.03534/) (2026)

## Abstract (arXiv)

Standard 3D Gaussian Splatting (3DGS) relies on known or pre-computed camera poses and a sparse point cloud, obtained from structure-from-motion (SfM) preprocessing, to initialize and grow 3D Gaussians. We propose a novel SfM-Free 3DGS (SFGS) method for video input, eliminating the need for known camera poses and SfM preprocessing. Our approach introduces a hierarchical training strategy that trains and merges multiple 3D Gaussian representations -- each optimized for specific scene regions -- into a single, unified 3DGS model representing the entire scene. To compensate for large camera motions, we leverage video frame interpolation models. Additionally, we incorporate multi-source supervision to reduce overfitting and enhance representation. Experimental results reveal that our approach significantly surpasses state-of-the-art SfM-free novel view synthesis methods. On the Tanks and Temples dataset, we improve PSNR by an average of 2.25dB, with a maximum gain of 3.72dB in the best scene. On the CO3D-V2 dataset, we achieve an average PSNR boost of 1.74dB, with a top gain of 3.90dB. The code is available at https://github.com/jibo27/3DGS_Hierarchical_Training.