In recent years, 3D Gaussian splatting (3DGS) has achieved remarkable progress in novel view synthesis. However, accurately reconstructing glossy surfaces under complex illumination remains challenging, particularly in scenes with strong specular reflections and multi-surface interreflections. To address this issue, we propose SSR-GS, a specular reflection modeling framework for glossy surface reconstruction. Specifically, we introduce a prefiltered Mip-Cubemap to model direct specular reflections efficiently, and propose an IndiASG module to capture indirect specular reflections. Furthermore, we design Visual Geometry Priors (VGP) that couple a reflection-aware visual prior via a reflection score (RS) to downweight the photometric loss contribution of reflection-dominated regions, with geometry priors derived from VGGT, including progressively decayed depth supervision and transformed normal constraints. Extensive experiments on both synthetic and real-world datasets demonstrate that SSR-GS achieves state-of-the-art performance in glossy surface reconstruction.
Figures & tables
Figure 1: Surface reconstruction results on toaster. CD denotes the Chamfer distance.
Figure 2: An overview of our SSR-GS pipeline. We rasterize 3D Gaussians to obtain surface buffers including normals, depth, opacity, diffuse component Cdiff , roughness, F0 , and alpha, and supervise them with geometry priors (GP). After rasterization, we extract a mesh via TSDF fusion and perform mesh-based ray tracing to estimate visibility wvis . Direct specular reflection is queried from a Mip-Cubemap environment map, while indirect specular reflection is modeled by IndiASG. Finally, we apply physically based deferred rendering (Eq. 8 ) to produce the rendered image. We additionally compute the visual prior (VP) (reflection score) and use it to down-weight reflection-dominated pixels in the photometric loss for stable geometry initialization.
ball
car
coffee
helmet
teapot
toaster
mean
Ref-NeRF [ 27 ]
1.55
14.93
12.24
29.48
9.23
42.87
18.38
ENVIDR [ 21 ]
0.74
7.10
9.23
1.66
2.47
6.45
4.61
UniSDF [ 28 ]
0.45
6.88
8.00
1.72
2.80
6.45
8.71
PGSR [ 2 ]
66.93
4.62
2.91
6.01
1.01
15.31
16.13
MILo [ 10 ]
61.30
4.47
2.36
4.87
0.62
16.49
15.02
GaussianShader [ 14 ]
7.03
14.05
14.93
9.33
7.17
13.08
10.93
Table 1: Quantitative surface reconstruction comparison on the ShinySynthetic dataset (normal MAE). CD is omitted since invisible inner surfaces in the GT meshes make it unreliable, as discussed in Ref-NeuS [ 8 ] for this dataset. Best results are highlighted in red and second-best in yellow.
angel
bell
cat
horse
luyu
potion
tbell
teapot
mean
CD ↓
PGSR [ 2 ]
0.77
3.08
3.39
0.90
1.16
3.99
4.59
4.57
2.81
MILo [ 10 ]
0.77
16.52
2.84
1.02
44.64
3.39
5.08
4.40
9.83
GaussianShader [ 14 ]
0.85
1.10
2.56
0.73
1.07
4.74
5.74
3.40
2.53
Ref-Gaussian [ 37 ]
0.45
0.70
1.68
0.64
0.88
0.81
0.59
1.01
0.95
Ref-GS [ 42 ]
0.41
0.74
1.73
0.47
0.89
1.05
0.52
0.88
0.84
Table 2: Quantitative surface reconstruction comparison on the GlossySynthetic dataset (CD ×102 and normal MAE). Best results are highlighted in red and second-best in yellow.
Mip-Cubemap
IndiASG
VGP
Stage 1
✓
✓
Stage 2
✓
✓
Table 3: Two-stage strategy.
Figure 3: Qualitative results of surface reconstruction on ShinySynthetic (car, coffee) and GlossySynthetic (cat, bell, teapot) datasets. Since the released GS-ROR 2 code cannot extract meshes on the ShinySynthetic, we instead visualize its normal results for comparison.
Figure 4: Qualitative results of surface reconstruction on Ref-Real (gardenspheres, toycar) dataset.
Mip-Cubemap
IndiASG
VGP
Full
w/o Mip
w/o Cube
w/o IASG
w/o VGP
w/o VP
w/o GP
w/o GP-D
w/o GP-N
Model
MAE ↓
1.56
1.62
1.52
2.25
1.55
2.15
1.53
2.86
1.52
MAE ↓
2.07
2.17
2.06
2.78
2.07
2.74
2.05
2.72
2.05
CD ↓
0.66
0.69
0.64
0.97
0.66
0.96
0.66
0.96
0.60
Table 4: Ablation study on different components. The first row corresponds to the ShinySynthetic dataset, while the remaining rows to the GlossySynthetic dataset.
The reflective appearance, especially strong and typically near-field specular reflections, poses a fundamental challenge for accurate surface reconstruction and novel view synthesis. Existing Gaussian splatting methods either fail to model near-field specular reflections or rely on explicit ray tracing at substantial computational cost. We present Ref-DGS, a reflective dual Gaussian splatting framework that addresses this trade-off by decoupling surface reconstruction from specular reflection within an efficient rasterization-based pipeline. Ref-DGS introduces a dual Gaussian scene representation consisting of geometry Gaussians and complementary local reflection Gaussians that capture near-field specular interactions without explicit ray tracing, along with a global environment reflection field for modeling far-field specular reflections. To predict specular radiance, we further propose a lightweight, physically-aware specular adaptive mixing shader that fuses global and local specular features. Experiments demonstrate that Ref-DGS achieves state-of-the-art performance on reflective scenes while training substantially faster than ray-based Gaussian methods.
Ningjing Fan, Yiqun Wang, Dong-Ming Yan +1
Chongqing University, China · MAIS, Institute of Automation, Chinese Academy of Sciences and UCAS, China · King Abdullah University of Science and Technology (KAUST), Saudi Arabia
Gaussian Splatting has significantly improved the quality of novel view synthesis with explicit Gaussian representation. However, we observed that existing 3D Gaussian Splatting methods (3DGS) often suffer from surface collapse issues on reflective regions, and thus produce inferior geometry and low-quality specular. In this work, we propose a physically-based deferred rendering framework, named Reflection-aware Gaussian Splatting (RGS), that can accurately model specular regions and improve novel view synthesis performance. Specifically, we found that a powerful 3D foundation model can provide a strong 3D geometric prior to foster correct geometric modeling. Based on this, we propose a cross-view shape consistency regularization to regularize the geometry surface with the large model prior and cross-view constraints. In this manner, our RGS can produce smoother geometric surfaces on reflective regions while reducing geometric hollows. To further improve rendering results on reflective regions, we present a reflection-aware densification strategy that is designed to capture specular variations across various views. With this strategy, our RGS is able to render novel views of objects in higher quality. Extensive experiments demonstrate our method consistently renders high-quality reflective objects, achieving state-of-the-art performance.
Xiaobiao Du, Yida Wang, Cheng Bi +2
University of Technology Sydney · Li Auto Inc. · Adelaide University
While novel view synthesis methods like Neural Radiance Fields (NeRF) and 3D Gaussian Splatting (3DGS) have shown impressive performance on common scenes, they often struggle with capturing view-dependent specular effects, particularly near-field reflections. Existing methods utilize environment maps, which only represent lights from infinity, or environment Gaussians, which rely solely on indirect supervision from reflections, are insufficient for accurately modeling such effects. Here, we propose a framework based on 2D Gaussian Splatting (2DGS) that leverages directly observable environment information in input images. Our approach also integrates a hybrid reflection appearance representation and view-dependent specular modeling. Experiments on public benchmarks and a newly curated near-field reflection dataset demonstrate that our method achieves superior rendering quality and geometric accuracy.
Yiming Liang, Tianyu Xiao, Hiroshi Ishikawa
Department of Computer Science and Communications Engineering, Waseda University, Japan