mlsanityData to 14 Oct 2026

RetinaGS: Scalable Training for Dense Scene Rendering with Billion-Scale 3D Gaussians

arXiv 2406.11836 ↗First posted 17 Jun 2024cs.CV cs.GRcited by 16

Not on the map. Why?

Abstract (arXiv)

In this work, we explore the possibility of training high-parameter 3D Gaussian splatting (3DGS) models on large-scale, high-resolution datasets. We design a general model parallel training method for 3DGS, named RetinaGS, which uses a proper rendering equation and can be applied to any scene and arbitrary distribution of Gaussian primitives. It enables us to explore the scaling behavior of 3DGS in terms of primitive numbers and training resolutions that were difficult to explore before and surpass previous state-of-the-art reconstruction quality. We observe a clear positive trend of increasing visual quality when increasing primitive numbers with our method. We also demonstrate the first attempt at training a 3DGS model with more than one billion primitives on the full MatrixCity dataset that attains a promising visual quality.

Key facts. Cited by 16 papers in the map (data through 14 Oct 2026). Builds on 32 of its 45 references that are in the map. First posted on arXiv 17 Jun 2024.

Builds on

32 of 45 references.

References · 32

  1. 2017Attention Is All You Needcited by 60,361
  2. 20233D Gaussian Splatting for Real-Time Radiance Field Renderingcited by 6,226
  3. 2022Instant Neural Graphics Primitives with a Multiresolution Hash Encodingcited by 3,723
  4. 2021Mip-NeRF 360: Unbounded Anti-Aliased Neural Radiance Fieldscited by 2,032
  5. 2021Mip-NeRF: A Multiscale Representation for Anti-Aliasing Neural Radiance Fieldscited by 1,964
  6. 2019Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelismcited by 1,868
  7. 2020GShard: Scaling Giant Models with Conditional Computation and Automatic Shardingcited by 1,631
  8. 2021Plenoxels: Radiance Fields without Neural Networkscited by 1,613
  9. 2019ZeRO: Memory Optimizations Toward Training Trillion Parameter Modelscited by 1,301
  10. 2022TensoRF: Tensorial Radiance Fieldscited by 1,280
  11. 2021Direct Voxel Grid Optimization: Super-fast Convergence for Radiance Fields Reconstructioncited by 964
  12. 2021PlenOctrees for Real-time Rendering of Neural Radiance Fieldscited by 905
20 more
  1. 2020NeRF++: Analyzing and Improving Neural Radiance Fieldscited by 861
  2. 2023ScanNet++: A High-Fidelity Dataset of 3D Indoor Scenescited by 764
  3. 2022Block-NeRF: Scalable Large Scene Neural View Synthesiscited by 724
  4. 2021Mega-NeRF: Scalable Construction of Large-Scale NeRFs for Virtual Fly-Throughscited by 398
  5. 2014Caffe: Convolutional Architecture for Fast Feature Embeddingcited by 288
  6. 2014One weird trick for parallelizing convolutional neural networkscited by 264
  7. 2021BungeeNeRF: Progressive Neural Radiance Field for Extreme Multi-scale Scene Renderingcited by 229
  8. 2023BakedSDF: Meshing Neural SDFs for Real-Time View Synthesiscited by 221
  9. 2023MERF: Memory-Efficient Radiance Fields for Real-time View Synthesis in Unbounded Scenescited by 218
  10. 2024VastGaussian: Vast 3D Gaussians for Large Scene Reconstructioncited by 209
  11. 2024CityGaussian: Real-time High-quality Large-Scale Scene Rendering with Gaussianscited by 192
  12. 2023MatrixCity: A Large-scale City Dataset for City-scale Neural Rendering and Beyondcited by 185
  13. 2023Delicate Textured Mesh Recovery from NeRF via Adaptive Surface Refinementcited by 106
  14. 2023Grid-guided Neural Radiance Fields for Large Urban Scenescited by 100
  15. 2022Compressing Volumetric Radiance Fields to 1 MBcited by 68
  16. 2022Masked Wavelet Representation for Compact Neural Radiance Fieldscited by 57
  17. 2023Efficient Large-scale Scene Representation with a Hybrid of High-resolution Grid and Plane Featurescited by 36
  18. 2015Towards Good Practices for Very Deep Two-Stream ConvNetscited by 25
  19. 2023City-on-Web: Real-time Neural Rendering of Large-scale Scenes on the Webcited by 14
  20. 2023Blockwise Parallel Transformer for Large Context Modelscited by 12

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 4; 2025: 8; 2026: 4.

16 papers.

Sort:

By area:

  1. 2024On Scaling Up 3D Gaussian Splatting Trainingcited by 49
  2. 2024Horizon-GS: Unified 3D Gaussian Splatting for Large-Scale Aerial-to-Ground Scenescited by 30
  3. 2024Momentum-GS: Momentum Gaussian Self-Distillation for High-Quality Large Scene Reconstructioncited by 28
  4. 2024Balanced 3DGS: Gaussian-wise Parallelism Rendering with Fine-Grained Tilingcited by 12
  5. 2025FlexGaussian: Flexible and Cost-Effective Training-Free Compression for 3D Gaussian Splattingcited by 11
  6. 2025GS-Scale: Unlocking Large-Scale 3D Gaussian Splatting Training via Host Offloadingcited by 5
  7. 2025Gaussian On-the-Fly Splatting: A Progressive Framework for Robust Near Real-Time 3DGS Optimizationcited by 4
  8. 2025Holistic Large-Scale Scene Reconstruction via Mixed Gaussian Splattingcited by 4
  9. 2025CLM: Removing the GPU Memory Barrier for 3D Gaussian Splattingcited by 3
  10. 2025A LoD of Gaussians: Unified Training and Rendering for Ultra-Large Scale Reconstruction with External Memorycited by 2
  11. 2026BlitzGS: City-Scale Gaussian Splatting at Lightning Speedcited by 2
  12. 2026TideGS: Scalable Training of Over One Billion 3D Gaussian Splatting Primitives via Out-of-Core Optimizationcited by 2
4 more
  1. 2026HiGS: A Hierarchical Rendering Architecture for Real-Time 3D Gaussian Splattingcited by 2
  2. 2025A-TDOM: Active TDOM via On-the-Fly 3DGScited by 1
  3. 2025LandMarkSystem Technical Reportcited by 0
  4. 2026TemporalGS: Training-Free Plug-and-Play Acceleration for 3D Gaussian Splatting Rendering via Temporal Priorscited by 0