cs.CVAug 14, 2026

HiCo-GS: Hierarchical Context Aggregation and Geometric Consistency for Octree Gaussian Splatting

Authors: Wei Zhang, Shengkai Yu, Shiqiang Gong, Qi Zhang, Qiang Li, Qi Wang

Organizations: School of Computer Science Northwestern Polytechnical University Xi’an, China · vivo BlueImage Lab vivo Mobile Communication Co., Ltd. Hangzhou, China

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Octree-based anchor Gaussian Splatting has emerged as a scalable representation for city-scale novel view synthesis, where multi-level anchors adaptively capture scene content from coarse building structures to fine architectural details. However, we identify a fundamental limitation in existing methods: cross-level feature isolation, where each level's anchor features are optimized independently with no inter-level communication, causing color drift on building facades and over-smoothing in textured regions. We present HiCo-GS, a high-fidelity reconstruction framework with two complementary modules. Cross-Level Context Aggregation (CLCA) enables bidirectional hierarchical prior injection by leveraging the octree's spatial containment structure to aggregate per-level context vectors into parent-self-child triplets, fused via a lightweight MLP with residual connection. Coarse-level structural priors flow down to inform fine-level anchors, while fine-level detail statistics feed back to prevent over-smoothing, at negligible computational overhead. Depth-Normal Geometric Consistency (DNGC) regularization enforces agreement between rendered normals and depth-derived normals through an alpha-weighted consistency loss, complemented by edge-aware smoothness losses with progressive warmup that exploit the strong planar priors ubiquitous in urban geometry to suppress floating artifacts. We further introduce the China-Pagoda dataset comprising 8 ancient Chinese pagodas with over 1,200 images each, featuring dense ornamental carvings, curved multi-layer eaves, and repetitive fine-grained textures. Extensive experiments on Mill19, UrbanScene3D, MatrixCity, and China-Pagoda demonstrate that HiCo-GS achieves state-of-the-art rendering quality and substantially cleaner geometry across real-world and synthetic urban benchmarks.Code: https://github.com/WZ-CS/HiCo-GS.

Figures & tables

Figure 1. HiCo-GS achieves high-fidelity urban reconstruction with cleaner geometry. Left: Mesh, rendered RGB, depth, and normal comparisons on Mill19 and UrbanScene3D between (a) HiCo-GS and (b) CityGS- X\mathcal{X} . Right: Representative scenes from our China-Pagoda dataset featuring extreme architectural complexity. Bottom: Radar charts summarizing quantitative results across both benchmarks. HiCo-GS (red) consistently outperforms prior methods.
Figure 2. Cross-level feature isolation and our solution. Top: In existing octree-based anchor Gaussian Splatting, multi-level anchors (L0 to LN) are optimized independently—each level’s features are decoded by shared MLPs with no inter-level communication. The t-SNE visualization ( right ) confirms this: features from all levels form a structureless mixture with inter-level cosine distances approaching 0.97, indicating near-orthogonal representations. Bottom: Our CLCA enables bidirectional hierarchical prior injection. Each anchor receives a context triplet from its parent (coarse), self, and child (fine) levels, fused through a lightweight MLP with residual connection. Coarse-level structural priors flow down while fine-level detail statistics feed back. After CLCA ( right ), features exhibit level-aware clustering with adjacent-level cosine distance reduced by up to 43%.
Figure 3. Overview of HiCo-GS. (a) A sparse point cloud is organized into a multi-level octree, where each anchor stores a feature vector, position, level, offsets, scale, and rotation. Visible anchors are enhanced by CLCA to produce f^\hat{\mathbf{f}} , decoded by three shared MLPs (conditioned on viewing direction and appearance) into neural Gaussians. Differentiable rasterization produces RGB image I\mathbf{I} , rendered normal nr\mathbf{n}_{r} , depth D\mathbf{D} (with derived normal nd\mathbf{n}_{d} ), distance d\mathbf{d} , and opacity α\boldsymbol{\alpha} . (b) CLCA performs spatially-aware level-wise aggregation to compute per-level context vectors, then assembles a parent-self-child triplet for each anchor via hierarchical context lookup. The triplet is fused by MLPcontext\mathrm{MLP}_{\text{context}} with a residual connection. (c) DNGC enforces consistency between nr\mathbf{n}_{r} (from Gaussian covariance) and nd\mathbf{n}_{d} (from depth) via Lnc\mathcal{L}_{\text{nc}} , and applies edge-aware smoothness losses Lns\mathcal{L}_{\text{ns}} , Lds\mathcal{L}_{\text{ds}} weighted by e−β∣∇I∣e^{-\beta|\nabla\mathbf{I}|} to enforce planarity while preserving sharp boundaries.
Figure 4. Quantitative analysis of CLCA on Building and Residence. (a, b) Change in inter-level cosine similarity after CLCA (warmer = larger increase). Both scenes show strong positive gains, particularly between fine levels (Building L5–L6: Δ=0.87\Delta{=}0.87 ; Residence L2–L3: Δ=0.77\Delta{=}0.77 ), confirming effective cross-level context flow. (c, d) Adjacent-level cosine distance before and after CLCA. On Building, all pairs show 11–88% reduction, with the finest pair (L5–L6) dropping from 0.97 to 0.12. On Residence, reductions range from 29% to 64% across all level pairs. More details are provided in the supplementary material.
Figure 5. Overview of the China-Pagoda dataset. More details are provided in the supplementary material.
MethodBuildingRubbleResidenceSci-Art
SSIM ↑\uparrowPSNR ↑\uparrowLPIPS ↓\downarrowSSIM ↑\uparrowPSNR ↑\uparrowLPIPS ↓\downarrowSSIM ↑\uparrowPSNR ↑\uparrowLPIPS ↓\downarrowSSIM ↑\uparrowPSNR ↑\uparrowLPIPS ↓\downarrow
Without geometric optimization
Mega-NeRF ( Turki et al., 2022 )0.54720.920.4540.55324.060.5080.62822.080.4010.77025.600.312
Switch-NeRF ( Lin et al., 2022 )0.57921.540.3970.56224.310.4780.65422.570.3520.79526.510.271
VastGaussian †\dagger ( Lin et al., 2024 )0.72821.800.2250.74225.200.2640.69921.010.2610.76122.640.261
3DGS ( Kerbl et al., 2023 )0.73822.530.2140.72525.510.3160.74522.360.2470.79124.130.262
Table 1. Quantitative comparison on Mill19 ( Turki et al., 2022 ) and UrbanScene3D ( Lin et al., 2022 ) . ↑\uparrow : higher is better, ↓\downarrow : lower is better. Best and second best results are highlighted. †\dagger : without decoupled appearance encoding.
DatasetScenesTotal imgsImgs/sceneTypeGeometry
Mill19 ( Turki et al., 2022 )2∼\sim 4K∼\sim 2KUrbanPlanar facades
UrbanScene3D ( Lin et al., 2022 )6∼\sim 5K∼\sim 0.8KUrbanPlanar facades
MatrixCity ( Li et al., 2023a )2∼\sim 10K∼\sim 5KSyntheticMixed
China-Pagoda (Ours)8∼\sim 10K+∼\sim 1.2K+HeritageExtreme
Table 2. Comparison with existing 3D reconstruction benchmarks. China-Pagoda provides dense per-scene coverage of geometrically extreme structures, complementing existing urban and cultural heritage datasets.
Figure 6. Qualitative results of our method and other methods in large-scale reconstruction datasets UrbanScene.
Figure 7. Qualitative results of ours and other methods in image and depth rendering on Mill-19 and Urbanscene3D datasets.
Figure 8. Qualitative mesh and texture comparison between CityGS- X\mathcal{X} and our method on the Residence.
MethodBaoantaBeitaDuobaofotaLingshan
SSIM ↑\uparrowPSNR ↑\uparrowLPIPS ↓\downarrowSSIM ↑\uparrowPSNR ↑\uparrowLPIPS ↓\downarrowSSIM ↑\uparrowPSNR ↑\uparrowLPIPS ↓\downarrowSSIM ↑\uparrowPSNR ↑\uparrowLPIPS ↓\downarrow
CityGS-v1 ( Liu et al., 2024a )0.74522.790.3690.75923.920.4270.80623.520.3080.87925.310.211
CityGS- X\mathcal{X} ( Gao et al., 2025 )0.89126.330.1680.77023.170.3500.81122.490.2170.95028.770.081
SplatCo ( Xiao et al., 2025 )0.82727.050.2730.64321.290.5780.79225.880.3040.55016.410.513
BlockGaussian ( Wu et al., 2025 )0.76023.480.3490.68621.740.4590.70020.290.4150.81625.060.269
Momentum-GS ( Fan et al., 2025 )0.81523.790.3590.71422.980.4620.79723.900.2570.92127.240.160
Table 3. Results on China-Pagoda benchmark. ↑\uparrow : higher is better, ↓\downarrow : lower is better.
MethodPSNR ↑\uparrowP ↑\uparrowR ↑\uparrowF1 ↑\uparrow
NeuS ( Wang et al., 2021 )16.76FAILFAILFAIL
Neuralangelo ( Li et al., 2023b )19.220.0800.0830.081
SuGaR ( Guédon and Lepetit, 2024 )OOMOOMOOMOOM
GOF ( Yu et al., 2024b )17.42FAILFAILFAIL
2DGS ( Huang et al., 2024 )21.350.2070.3900.270
CityGS ( Liu et al., 2024a )27.460.3620.6370.462
Table 4. Results on MatrixCity ( Li et al., 2023a ) . Best and second best results are highlighted.
CLCADNGCBuild.Rub.Res.Sci.Avg. Time
––22.4020.5721.8122.042h44min
Mean–22.5126.2924.3726.432h51min
–✓22.6625.7523.2424.392h44min
Mean✓22.6728.0924.4126.642h51min
Max/Attn✓22.6027.9724.3026.532h52min
Table 5. Ablation study. PSNR (dB) ↑\uparrow on Mill19/UrbanScene3D.
Figure 1. Cross-level feature analysis on the Building scene (8 octree levels, 5.98M anchors). (a) Baseline cosine distance matrix: all off-diagonal entries fall within 0.94–1.00, with incoherence worsening at finer levels (L6 ↔\leftrightarrow L7: 0.999). (b) After CLCA: a block-diagonal structure emerges, with fine-level coherence dramatically improved (L5 ↔\leftrightarrow L6: 0.995 →\to 0.343). (c) Adjacent-level distance reductions concentrate on the finest pairs (L5 ↔\leftrightarrow L6: −- 65.5%, L6 ↔\leftrightarrow L7: −- 25.0%), while coarse pairs change by ≤\leq 4%. (d) Per-pair PCA: baseline features (top) are diffusely mixed across levels; after CLCA (bottom), fine-level pairs form clearly separated clusters with increased PC1 explained variance, indicating emergent scale-discriminative structure.
Figure 2. Cross-level feature analysis on the Rubble scene (5 octree levels). The same analysis protocol as Fig. 1 . The baseline exhibits uniform near-orthogonality across all level pairs. After CLCA, the finest pair (L3 ↔\leftrightarrow L4) shows a 28.8% distance reduction, while coarse pairs remain stable. Per-pair PCA confirms that CLCA selectively organizes fine-level features into separable clusters without disrupting coarse-level representations.
Figure 3. Cross-level feature analysis on the Residence scene (6 octree levels). CLCA reduces adjacent-level distances by up to 40.1% at the finest pair (L4 ↔\leftrightarrow L5), with coarse pairs showing minimal change. The adaptive coupling pattern is consistent with the other scenes.
Figure 4. Cross-level feature analysis on the SciArt scene (7 octree levels). SciArt exhibits the largest fine-level reductions among all scenes (L5 ↔\leftrightarrow L6: −- 62.9%), consistent with its higher geometric complexity requiring stronger cross-level communication.
Figure 5. t-SNE visualization of anchor features across all octree levels. Each panel shows the baseline ( left ) and after CLCA ( right ) for one scene, with stratified sampling ensuring equal representation per level. In the baseline, all levels form a single unstructured mixture. After CLCA, fine-level features consistently separate into distinct peripheral clusters while coarse levels share a common core, demonstrating the emergent hierarchical specialization induced by cross-level context aggregation.
Figure 6. DNGC ablation on the Building scene. Top: With DNGC, rendered normals exhibit clean, consistent orientations on planar surfaces such as rooftops and walls, and depth maps show smooth gradients with sharp object boundaries. Bottom: Without DNGC, normals are noticeably noisier with inconsistent directions on the same regions, and depth maps contain discontinuities and floating artifacts around building facades and ground planes.
Figure 8. China-Pagoda sample images (Part 1): Beita, Duobaota, Lingshan, and Liuhe pagodas shown from multiple drone viewpoints. These pagodas feature octagonal and hexagonal cross-sections, tiered eaves with hanging bells, intricate brick carvings, and arched doorways.
Figure 9. China-Pagoda sample images (Part 2): Yunjusi, Yongwangta, Yuhuangta, and Baoanta pagodas. These structures exhibit heavily weathered surfaces, dense repetitive textures from layered brick construction, and fine-grained ornamental details at varying scales.
Figure 10. Reconstruction of Duobaota. From left to right: textured rendering, extracted mesh, and top-down views. The multi-layer eaves, hanging bells, and hexagonal base geometry are faithfully reconstructed with clean surface normals and sharp edge boundaries.
Figure 12. Reconstruction of Lingshan. This pagoda features prominent curved eaves with decorative bells and a surrounding stone balustrade. The mesh accurately captures eave curvature, the octagonal tiered structure, and fine railing geometry at the base.
Figure 14. Reconstruction of Beita. The dense brick carvings and weathered surface patterns are well preserved in both the rendered appearance and the extracted mesh geometry.
Figure 16. Reconstruction of Yongwangta. Clean mesh extraction captures the tapered silhouette and tiered eave overhangs characteristic of this architectural style.
Figure 18. Full-scene comparison on Residence. Top: HiCo-GS produces consistent facade coloring and clean planar surfaces on rooftops and roads. Bottom: CityGS- X\mathcal{X} exhibits more floating artifacts and noisier mesh geometry, especially around vegetation boundaries and overpass structures.

Explore similar work

CardsList
Aug 20, 2026cs.CV

TopoGS: Topology-Aware Anchor Feature Aggregation for Large-Scale 3D Gaussian Splatting

Octree-based 3D Gaussian Splatting organizes anchors into multi-level hierarchies for level-of-detail rendering, but features at different levels are typically optimized independently, leaving the octree topology underused during feature learning. We observe that uniform cross-level aggregation produces asymmetric effects: fine-level anchors benefit from coarse context, whereas coarse-level anchors require selective information from their descendants. We therefore propose TopoGS, a topology-aware anchor feature aggregation framework with two lightweight components. Hierarchical Anchor Coupling establishes bidirectional cross-level gradient pathways by fusing per-level context triplets with a residual MLP. Structure-Aware Containment Aggregation uses octree containment and hash-based matching to distinguish anchors with valid parent-child relations from isolated anchors, then applies soft weighting to accommodate varying topological sparsity. Experiments on ten scenes from Mill19, UrbanScene3D, Tanks & Temples, MatrixCity, and WHU show consistent improvements over state-of-the-art methods. TopoGS achieves average PSNR gains of 2.13, 1.78, and 0.29 dB over the strongest reported baseline on aerial, ground-level, and synthetic-cartographic scenes, respectively, while rendering faster and using less memory. Code is available at https://github.com/WZ-CS/TopoGS.

Wei Zhang, Shiqiang Gong, Shengkai Yu +4

School of Computer Science, and with the School of Artificial Intelligence, Optics and Electronics (iOPEN), Northwestern Polytechnical University, Xi’an 710072, P. R. China · School of Computer Science, Northwestern Polytechnical University, Xi’an 710072, P. R. China · School of Artificial Intelligence, Optics and Electronics (iOPEN), Northwestern Polytechnical University, Xi’an 710072, P. R. China +1

3D Gaussian SplattingMulti-Scale Feature Fusion
AskPDF
May 13, 2026cs.GR

BlitzGS: City-Scale Gaussian Splatting at Lightning Speed

We present BlitzGS, a distributed 3DGS framework that reduces active Gaussian workload for fast city-scale reconstruction. BlitzGS manages this workload at three coupled levels. At the system level, the framework shards Gaussians across GPUs by index parity rather than spatial blocks. This approach mitigates the cross-block visibility redundancy inherent in spatial partitioning. Furthermore, it distributes each rendering step through a single cross-GPU exchange that routes projected Gaussians to their tile owners. At the model level, scheduled importance-scoring passes shrink the global Gaussian population. During these passes, the framework generates a per-Gaussian visibility weight to bias density-control updates toward contributing primitives and a per-view importance mask for the view-level renderer. At the view level, BlitzGS trims each camera's active set with a distance-based LOD gate to exclude excessively fine primitives for the current frustum and the importance-based culling mask to skip Gaussians with negligible cross-view contribution. On large-scale benchmarks, BlitzGS matches the rendering quality of recent large-scale baselines while delivering an order-of-magnitude speedup, training city-scale scenes in tens of minutes. Our code is available at https: //github.com/AkierRaee/BlitzGS.

Zhongtao Wang, Huishan Au, Yilong Li +6

Peking University, China

3D Gaussian Splatting3D Gaussian Pruning3DGS AccelerationGPU Acceleration3D ReconstructionGaussian Splatting
AskPDF
May 29, 2026cs.CV

HiGS: A Hierarchical Rendering Architecture for Real-Time 3D Gaussian Splatting

3D Gaussian Splatting (3DGS) has become the standard for real-time novel view synthesis on commodity GPUs. Its pipeline ties spatial partitioning and rasterization to one tile size, yet the two pull in opposite directions: partitioning, which bins and depth-sorts gaussians, grows cheaper with larger tiles, while rasterization gets cheaper with smaller ones. Prior acceleration work reduces the cost of individual stages but keeps both locked to that single scale, where a few dense tiles dominate frame time. We present Hierarchically Tiled Gaussian Splatting (HiGS), which gives each its own scale: partitioning runs over coarse macro-tiles, while rasterization runs over the fine render tiles within them. Rasterization work is then issued in proportion to the gaussians in each macro-tile rather than per tile, so dense regions spread across many parallel units instead of serializing through one. Across tested scenes, HiGS renders up to 15.8x faster than the original 3DGS and outperforms every other rasterizer we evaluate, while preserving exact front-to-back alpha compositing.

Dawid Pająk, Martin Bisson, Rodolfo Lima

NVIDIA

3D Gaussian Splatting3DGS AccelerationGPU AccelerationGaussian Splatting
AskPDF