Efficient multi-view training for 3D Gaussian Splatting

• URL: http://arxiv.org/abs/2506.12727v2
• Date: Tue, 17 Jun 2025 02:12:36 GMT
• Title: Efficient multi-view training for 3D Gaussian Splatting
• Authors: Minhyuk Choi, Injae Kim, Hyunwoo J. Kim,
• Abstract summary: 3D Gaussianting (3DGS) has emerged as a preferred choice alongside Neural Radiance Fields (NeRF) in inverse rendering due to its superior speed.<n>Currently, the common approach in 3DGS is to utilize "single-view" mini-batch training, where only one image is processed per iteration.<n>We observe that such single-view training can lead to suboptimal optimization due to increased variance in mini-batch gradients, highlighting the necessity for multi-view training.
• Score: 13.533415744533995
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: 3D Gaussian Splatting (3DGS) has emerged as a preferred choice alongside Neural Radiance Fields (NeRF) in inverse rendering due to its superior rendering speed. Currently, the common approach in 3DGS is to utilize "single-view" mini-batch training, where only one image is processed per iteration, in contrast to NeRF's "multi-view" mini-batch training, which leverages multiple images. We observe that such single-view training can lead to suboptimal optimization due to increased variance in mini-batch stochastic gradients, highlighting the necessity for multi-view training. However, implementing multi-view training in 3DGS poses challenges. Simply rendering multiple images per iteration incurs considerable overhead and may result in suboptimal Gaussian densification due to its reliance on single-view assumptions. To address these issues, we modify the rasterization process to minimize the overhead associated with multi-view training and propose a 3D distance-aware D-SSIM loss and multi-view adaptive density control that better suits multi-view scenarios. Our experiments demonstrate that the proposed methods significantly enhance the performance of 3DGS and its variants, freeing 3DGS from the constraints of single-view training.

Related papers

• 3DGabSplat: 3D Gabor Splatting for Frequency-adaptive Radiance Field Rendering [50.04967868036964]
  3D Gaussian Splatting (3DGS) has enabled real-time rendering while maintaining high-fidelity novel view synthesis.<n>We propose 3D Gabor Splatting (3DGabSplat) that incorporates a novel 3D Gabor-based primitive with multiple directional 3D frequency responses.<n>We achieve 1.35 dBR gain over 3D with simultaneously reduced number of primitive memory consumption.
  arXiv  Detail & Related papers  (2025-08-07T12:49:44Z)
• FlexGS: Train Once, Deploy Everywhere with Many-in-One Flexible 3D Gaussian Splatting [57.97160965244424]
  3D Gaussian splatting (3DGS) has enabled various applications in 3D scene representation and novel view synthesis.<n>Previous approaches have focused on pruning less important Gaussians, effectively compressing 3DGS.<n>We present an elastic inference method for 3DGS, achieving substantial rendering performance without additional fine-tuning.
  arXiv  Detail & Related papers  (2025-06-04T17:17:57Z)
• EVolSplat: Efficient Volume-based Gaussian Splatting for Urban View Synthesis [61.1662426227688]
  Existing NeRF and 3DGS-based methods show promising results in achieving photorealistic renderings but require slow, per-scene optimization.<n>We introduce EVolSplat, an efficient 3D Gaussian Splatting model for urban scenes that works in a feed-forward manner.
  arXiv  Detail & Related papers  (2025-03-26T02:47:27Z)
• LLGS: Unsupervised Gaussian Splatting for Image Enhancement and Reconstruction in Pure Dark Environment [18.85235185556243]
  We propose an unsupervised multi-view stereoscopic system based on 3D Gaussian Splatting.<n>This system aims to enhance images in low-light environments while reconstructing the scene.<n> Experiments conducted on real-world datasets demonstrate that our system outperforms state-of-the-art methods in both low-light enhancement and 3D Gaussian Splatting.
  arXiv  Detail & Related papers  (2025-03-24T13:05:05Z)
• Gaussian On-the-Fly Splatting: A Progressive Framework for Robust Near Real-Time 3DGS Optimization [8.422116335889163]
  3D Gaussian Splatting (3DGS) achieves high-fidelity rendering with fast real-time performance.<n>Existing methods rely on offline training after full Structure-from-Motion (SfM) processing.<n>We introduce On-the-Fly GS, a progressive framework enabling near real-time 3DGS optimization during image capture.
  arXiv  Detail & Related papers  (2025-03-17T11:47:58Z)
• HuGDiffusion: Generalizable Single-Image Human Rendering via 3D Gaussian Diffusion [50.02316409061741]
  HuGDiffusion is a learning pipeline to achieve novel view synthesis (NVS) of human characters from single-view input images.<n>We aim to generate the set of 3DGS attributes via a diffusion-based framework conditioned on human priors extracted from a single image.<n>Our HuGDiffusion shows significant performance improvements over the state-of-the-art methods.
  arXiv  Detail & Related papers  (2025-01-25T01:00:33Z)
• SplatFormer: Point Transformer for Robust 3D Gaussian Splatting [18.911307036504827]
  3D Gaussian Splatting (3DGS) has recently transformed photorealistic reconstruction, achieving high visual fidelity and real-time performance.<n> rendering quality significantly deteriorates when test views deviate from the camera angles used during training, posing a major challenge for applications in immersive free-viewpoint rendering and navigation.<n>We introduce SplatFormer, the first point transformer model specifically designed to operate on Gaussian splats.<n>Our model significantly improves rendering quality under extreme novel views, achieving state-of-the-art performance in these challenging scenarios and outperforming various 3DGS regularization techniques, multi-scene models tailored for sparse view synthesis, and diffusion
  arXiv  Detail & Related papers  (2024-11-10T08:23:27Z)
• MCGS: Multiview Consistency Enhancement for Sparse-View 3D Gaussian Radiance Fields [73.49548565633123]
  Radiance fields represented by 3D Gaussians excel at synthesizing novel views, offering both high training efficiency and fast rendering. Existing methods often incorporate depth priors from dense estimation networks but overlook the inherent multi-view consistency in input images. We propose a view framework based on 3D Gaussian Splatting, named MCGS, enabling scene reconstruction from sparse input views.
  arXiv  Detail & Related papers  (2024-10-15T08:39:05Z)
• MVGS: Multi-view-regulated Gaussian Splatting for Novel View Synthesis [22.80370814838661]
  Recent works in volume rendering, textite.g. NeRF and 3D Gaussian Splatting (3DGS), significantly advance the rendering quality and efficiency. We propose a new 3DGS optimization method embodying four key novel contributions.
  arXiv  Detail & Related papers  (2024-10-02T23:48:31Z)
• WE-GS: An In-the-wild Efficient 3D Gaussian Representation for Unconstrained Photo Collections [8.261637198675151]
  Novel View Synthesis (NVS) from unconstrained photo collections is challenging in computer graphics. We propose an efficient point-based differentiable rendering framework for scene reconstruction from photo collections. Our approach outperforms existing approaches on the rendering quality of novel view and appearance synthesis with high converge and rendering speed.
  arXiv  Detail & Related papers  (2024-06-04T15:17:37Z)
• LP-3DGS: Learning to Prune 3D Gaussian Splatting [71.97762528812187]
  We propose learning-to-prune 3DGS, where a trainable binary mask is applied to the importance score that can find optimal pruning ratio automatically. Experiments have shown that LP-3DGS consistently produces a good balance that is both efficient and high quality.
  arXiv  Detail & Related papers  (2024-05-29T05:58:34Z)
• MVSplat: Efficient 3D Gaussian Splatting from Sparse Multi-View Images [102.7646120414055]
  We introduce MVSplat, an efficient model that, given sparse multi-view images as input, predicts clean feed-forward 3D Gaussians. On the large-scale RealEstate10K and ACID benchmarks, MVSplat achieves state-of-the-art performance with the fastest feed-forward inference speed (22fps)
  arXiv  Detail & Related papers  (2024-03-21T17:59:58Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.