← Back to all news
arXiv Computer Vision

Compact Feed-Forward 3D Gaussians via Saliency-Guided Primitive Merging

The paper introduces a structure‑aware merging pipeline that consolidates per‑pixel 3D Gaussian primitives from any feed‑forward reconstruction method into a compact, content‑adaptive Gaussian set. By grouping spatially coherent Gaussians with adaptive superpixel segmentation guided by a saliency map, compressing clusters via a learned encoder, and merging representations across views using geometric overlap and feature similarity, the method reduces the number of Gaussians to about one‑twentieth of the original while preserving visual quality. A level‑of‑detail decoder allows controllable resolution, and the pipeline operates as a backbone‑agnostic post‑processing module, improving robustness and rendering efficiency.

Related stories

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
efficiency
More like this →
arXiv AI
Sep 10

Adaptive Densification for High-Fidelity and Efficient Sparse Gaussian Splatting in Arbitrary-Scale Super-Resolution

The paper introduces QuADA-GS, a method for Arbitrary-Scale Super-Resolution that dynamically densifies 2D Gaussian splatting based on low‑resolution input. By allocating Gaussians adaptively to structurally complex regions and employing a sparse communication mechanism, it balances high visual fidelity with lower computational cost. Experiments show that this approach achieves a competitive trade‑off between quality and efficiency for super‑resolution tasks.

By Giulio Federico, Giuseppe Amato, Claudio Gennaro, Fabio Carrara, Marco Di Benedetto
multimodal
More like this →
arXiv AI
Sep 28

Spackle: Completing Large View Single Image NVS with Adaptive Gaussians

Spackle is a lightweight residual learning framework designed to improve large-view single-image novel view synthesis (NVS) by mitigating capacity competition in hybrid decoupled systems that combine 3D Gaussian Splatting (3DGS) and diffusion models. It operates in three stages: predicting base 3DGS attributes, automatically identifying poorly reconstructed regions, and learning a residual 3DGS focused on those areas. During inference, Spackle merges the baseline and augmented Gaussians to produce high-fidelity novel views, achieving state‑of‑the‑art performance on large-view-deviation cases.

By Xuanzhi Liu, Yuhe Zhou, Xinyi Wu, Zhenyao Wu, Jinghao Chen, Ruize Han, Song Wang
diffusionbenchmarks
More like this →
Hugging Face Trending Papers
Sep 17

PhGS: Post-Hoc Pruning and Refinement of Single-View Feed-Forward 3D Gaussian Reconstructions

The paper introduces PhGS, a post‑hoc pruning and refinement pipeline for single‑view feed‑forward 3D Gaussian Splatting models. By keeping the base network frozen, it applies importance‑score‑based pruning followed by a lightweight recurrent refinement module to reduce spatial redundancy while restoring image quality. The method is backbone‑agnostic, integrates seamlessly with existing baselines, preserves novel‑view rendering fidelity, achieves significant memory reduction, and allows flexible inference‑time keep ratios.

efficiency
More like this →
arXiv Computer Vision
Sep 18

PhGS: Post-Hoc Pruning and Refinement of Single-View Feed-Forward 3D Gaussian Reconstructions

The paper introduces PhGS, a post‑hoc pruning and refinement pipeline for single‑view feed‑forward 3D Gaussian Splatting models. It keeps the base network frozen and applies importance‑score‑based pruning followed by a lightweight recurrent refinement module to reduce spatial redundancy while maintaining rendering quality. The method is backbone‑agnostic, integrates seamlessly with existing baselines, and allows flexible inference‑time keep ratios for different application needs.

By Rinto Yagawa, Han Cheng, Dieter Schmalstieg, Hideo Saito, Shohei Mori
efficiency
More like this →
arXiv Computer Vision
Aug 28

ReSplat: Learning Recurrent Gaussian Splatting

ReSplat introduces a recurrent Gaussian splatting model that iteratively refines 3D Gaussians using the rendering error as a feedback signal, avoiding explicit gradient computation. The method starts from a compact reconstruction in a subsampled space, producing far fewer Gaussians than prior per‑pixel models, which reduces computational cost. Experiments on multiple datasets, view counts, and resolutions show state‑of‑the‑art performance with faster rendering speeds.

By Haofei Xu, Daniel Barath, Andreas Geiger, Marc Pollefeys
benchmarks
More like this →