Lune

ICLR2025Top-tier venue

GaussianBlock: Building Part-Aware Compositional and Editable 3D Scene by Primitives and Gaussians

Shuyi Jiang, Qihao Zhao, Hossein Rahmani, De Wen Soh, Jun Liu, Na Zhao

Abstract

Recently, with the development of Neural Radiance Fields and Gaussian Splatting, 3D reconstruction techniques have achieved remarkably high fidelity. However, the latent representations learned by these methods are highly entangled and lack interpretability. In this paper, we propose a novel part-aware compositional reconstruction method, called GaussianBlock, that enables semantically coherent and disentangled representations, allowing for precise and physical editing akin to building blocks, while simultaneously maintaining high fidelity. Our GaussianBlock introduces a hybrid representation that leverages the advantages of both primitives, known for their flexible actionability and editability, and 3D Gaussians, which excel in reconstruction quality. Specifically, we achieve semantically coherent primitives through a novel attention-guided centering loss derived from 2D semantic priors, complemented by a dynamic splitting and fusion strategy. Furthermore, we utilize 3D Gaussians that hybridize with primitives to refine structural details and enhance fidelity. Additionally, a binding inheritance strategy is employed to strengthen and maintain the connection between the two. Our reconstructed scenes are evidenced to be disentangled, compositional, and compact across diverse benchmarks, enabling seamless, direct and precise editing while maintaining high quality. The code is available at https://github.com/Jiangshuyi0V0/GaussianBlock.

Ask Lune

Know this paper in minutes, not an afternoon.

Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.

5 top-tier papers have cited it since. Know what they found before your reviewers do.

Questions to start from

Lune can draw on

the full paper · the top-tier papers that cite it · related work

Get the answer

Free to start. No credit card required.

Lineage

  • D-Prism: Differentiable Primitives for Structured Dynamic Modeling cites 3D Gaussian Splatting for Real-Time Radiance Field Rendering
  • D-Prism: Differentiable Primitives for Structured Dynamic Modeling cites Self-Supervised Learning of Hybrid Part-Aware 3D Representations of 2D Gaussians and Superquadrics
  • Nerfstudio: A Modular Framework for Neural Radiance Field Development cites Mip-NeRF: A Multiscale Representation for Anti-Aliasing Neural Radiance Fields
  • Nerfstudio: A Modular Framework for Neural Radiance Field Development cites Mip-NeRF 360: Unbounded Anti-Aliased Neural Radiance Fields
  • DualPrim: Compact 3D Reconstruction with Positive and Negative Primitives cites 3D Gaussian Splatting for Real-Time Radiance Field Rendering
  • 3D Gaussian Splatting for Real-Time Radiance Field Rendering cites Mip-NeRF: A Multiscale Representation for Anti-Aliasing Neural Radiance Fields
  • Self-Supervised Learning of Hybrid Part-Aware 3D Representations of 2D Gaussians and Superquadrics cites Segment Anything
  • Self-Supervised Learning of Hybrid Part-Aware 3D Representations of 2D Gaussians and Superquadrics cites 3D Gaussian Splatting for Real-Time Radiance Field Rendering
Builds onCited byLeft to right by year; the more cited, the higher.

Cited by top-tier papers5

  1. Benchmarking PhD-Level Coding in 3D Geometric Computer VisionWenyi Li, Renkai Luo, Yue Yu, Huan-ang Gao et al.CVPR 2026 · 2 citations
  2. Self-Supervised Learning of Hybrid Part-Aware 3D Representations of 2D Gaussians and SuperquadricsZhirui Gao, Renjiao Yi, Yuhang Huang, Wei Chen et al.ICCV 2025 · 2 citations
  3. Graph Smoothing for Enhanced Local Geometry Learning in Point Cloud AnalysisShangbo Yuan, Jie Xu, Ping Hu, Xiaofeng Zhu et al.AAAI 2026
  4. D-Prism: Differentiable Primitives for Structured Dynamic ModelingXingyuan Yu, Yijin Li, Chong Zeng, Yuhang Ming et al.CVPR 2026
  5. DualPrim: Compact 3D Reconstruction with Positive and Negative PrimitivesXiaoxu Meng, Zhongmin Chen, Bo Yang, Weikai Chen et al.CVPR 2026

Builds on22

  1. Segment AnythingAlexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao et al.ICCV 2023 · 13,211 citations
  2. 3D Gaussian Splatting for Real-Time Radiance Field RenderingBernhard Kerbl, Georgios Kopanas, Thomas Leimkühler, George DrettakisSIGGRAPH 2023 · 5,687 citations
  3. Mip-NeRF: A Multiscale Representation for Anti-Aliasing Neural Radiance FieldsJonathan T. Barron, Ben Mildenhall, Matthew Tancik, Peter Hedman et al.ICCV 2021 · 2,700 citations
  4. Mip-NeRF 360: Unbounded Anti-Aliased Neural Radiance FieldsJonathan T. Barron, Ben Mildenhall, Dor Verbin, Pratul P. Srinivasan et al.CVPR 2022 · 1,603 citations
  5. Nerfstudio: A Modular Framework for Neural Radiance Field DevelopmentMatthew Tancik, Ethan Weber, Evonne Ng, Ruilong Li et al.SIGGRAPH 2023 · 592 citations

Related papers

  1. TextSplat: Text-Guided Semantic Fusion for Generalizable Gaussian SplattingZhicong Wu, Hongbin Xu, Gang Xu, Ping Nie et al.ACM MM 2025 · 6 citations
  2. Prune Wisely, Reconstruct Sharply: Compact 3D Gaussian Splatting via Adaptive Pruning and Difference-of-Gaussian PrimitivesHaoran Wang, Guoxi Huang, Fan Zhang, David Bull et al.CVPR 2026 · 6 citations
  3. GSRecon: Efficient Generalizable Gaussian Splatting for Surface Reconstruction from Sparse ViewsHang Yang, Le Hui, Jianjun Qian, Jin Xie et al.ICCV 2025 · 1 citation
  4. 3D Geometry-aware Deformable Gaussian Splatting for Dynamic View SynthesisZhicheng Lu, Xiang Guo, Le Hui, Tianrui Chen et al.CVPR 2024 · 33 citations
  5. Holistic Large-Scale Scene Reconstruction via Mixed Gaussian SplattingChuandong Liu, Huijiao Wang, Lei Yu, Gui-Song XiaNeurIPS 2025 · 4 citations
Dusk over the sea between two cliffs drawn in fine vertical lines