arXiv Computer Vision

EffGS: Efficient and High-Fidelity Gaussian Splatting

arXiv Computer Vision
Sep 4

F4Splat: Feed-Forward Predictive Densification for Feed-Forward 3D Gaussian Splatting

F4Splat introduces a feed‑forward predictive densification strategy for 3D Gaussian splatting that allocates Gaussians based on a densification‑score guided by spatial complexity and multi‑view overlap. The method predicts per‑region scores to estimate required Gaussian density, enabling explicit control over the total Gaussian budget without retraining. This adaptive allocation reduces redundancy in simple regions and minimizes duplicate Gaussians across overlapping views, yielding compact yet high‑quality 3D representations and superior novel‑view synthesis performance with fewer Gaussians.

By Injae Kim, Chaehyeon Kim, Minseong Bae, Minseok Joo, Hyunwoo J. Kim
arXiv Computer Vision
Oct 12

CompactSplat: Spatially Adaptive Gaussian Distribution for Feedforward Scene Reconstruction

CompactSplat introduces a feedforward scene reconstruction method that adaptively allocates 3D Gaussian primitives based on texture-guided spatial partitioning and hierarchical sampling. By concentrating dense Gaussians in complex regions and using expanded primitives for flat areas, it reduces redundancy and better captures planar geometry. The framework also allows dynamic sparsity control during inference, enabling trade-offs between rendering quality and memory usage without retraining.

By Bing He, Jingnan Gao, Yunuo Chen, Zhengxue Cheng, Rong Xie, Li Song, Wenjun Zhang
arXiv Computer Vision
Sep 3

InceptionGS: Generative Bootstrapping for Large-Scale Gaussian Splatting under Unstructured View Sampling

InceptionGS is a method that improves large‑scale Gaussian splatting for scenes captured with unstructured view sampling. It starts from an initial Gaussian splatting and selectively repairs areas affected by sparse views by integrating scene‑ and view‑adaptive generative priors, while keeping well‑covered regions unchanged. Experiments on real‑world scenes show that this hybrid reconstruction‑generation approach yields higher‑fidelity results than existing methods.

By Tianheng Lu, Guangyu Wang, Ruqi Huang, Lu Fang
arXiv Computer Vision
Sep 28

Gauss What You Need: Compact Gaussian Splatting Across Scene Scales

Gauss What You Need: Compact Gaussian Splatting Across Scene Scales introduces TangoGS, a method that automatically selects the number of Gaussian primitives for 3D Gaussian Splatting by combining capture-derived model sizing with training-based adaptation. The approach first estimates a learning allowance based on the capture’s total pixels, then adjusts the number of Gaussians during training according to reconstruction quality. On standard benchmarks, TangoGS matches the best baseline’s PSNR while using 48% fewer Gaussians, and on larger captures it scales automatically to achieve the highest mean PSNR with 2.3× more Gaussians.

By Afif Boudaoud, Jiayi Liu, Alexandru Calotoiu, Torsten Hoefler
arXiv Computer Vision
Nov 5

BudgetSplat: Budget-Controllable Generalizable Gaussian Splatting

BudgetSplat introduces a budget‑controllable framework for 3D Gaussian splatting that adapts the final Gaussian representation to a user‑specified preservation ratio. It jointly optimizes importance estimation, spatial coverage, and cross‑view redundancy using Intra‑Cross Mask Supervision and budget‑conditioned feature modulation, enabling a single trained model to support varying numbers of context views and multiple test‑time budgets without per‑budget retraining. Experiments on indoor reconstruction and cross‑dataset generalization show consistent quality‑size trade‑offs, significantly reducing representation storage and rendering cost while maintaining rendering quality.

By Yuxi Hu, Jun Zhang, Wei Mao, Yuchen Rao, Friedrich Fraundorfer
Hugging Face Trending Papers
Nov 2

OctreeSplat: Hierarchical Octree-Guided Feed-Forward Gaussian Splatting

OctreeSplat introduces a hierarchical octree-guided feed‑forward Gaussian splatting method that extracts multi‑view features and geometric priors from a frozen visual foundation model. By constructing a hierarchical octree and predicting densification scores, it allocates Gaussian primitives at spatially varying densities, improving depth rendering and surface reconstruction. The approach requires only 76 million additional parameters and 35 GPU‑hours to train, and demonstrates competitive RGB view synthesis, depth synthesis, and surface reconstruction on Replica, TUM, Tanks and Temples, and DTU datasets while reducing the number of Gaussian primitives needed.

arXiv AI
Sep 4

TruncGradGS: Improved 3D Gaussian Splatting via Truncated Gradient Updates

The paper introduces TruncGradGS, a piecewise truncated gradient approach that mitigates gradient vanishing in 3D Gaussian Splatting, enhancing optimization stability and robustness to initializations. It demonstrates consistent improvements over random and COLMAP initializations in both static and dynamic settings. Additionally, the authors highlight limitations of existing dynamic scene benchmarks and present a new synthetic dataset for evaluating dynamic Gaussian Splatting.

By Theo Morales, Nhat-Quynh Le-Pham, Robin Atkins, Binh-Son Hua
arXiv Computer Vision
Oct 22

Unfolding 3D Gaussian Splatting via Iterative Gaussian Synopsis

The paper introduces Iterative Gaussian Synopsis, a top‑down unfolding framework that generates compact, progressive levels of detail for 3D Gaussian Splatting (3DGS). Starting from a full‑resolution 3DGS model, it iteratively prunes Gaussian primitives using a learnable mask, building a multi‑level hierarchy that preserves visual quality while reducing storage. The method combines hierarchical spatial grids with a shared Anchor Codebook to capture global structure and local details, enabling efficient, level‑specific rendering with minimal data overhead.

By Yuqin Lu, Yihua Dai, Yang Zhou, Guiqing Li, Shengfeng He