- Rating
- 3(+1)
- Confidence
- 4(+1)
- Informativeness
- 2(+1)
- Presentation
- 2
- Soundness
- 2
Read review
Paper summary
This paper proposes SAS-LoD, which shares Gaussian positions, rotations, and spherical harmonic coefficients across levels while introducing level-specific scale and opacity residuals. Codebook compression and on-demand decoding further reduce resident memory. Experiments on six indoor and outdoor scenes evaluate rendering quality, throughput, and memory.
Critical strengths
- A clear representation design. Scale and opacity adaptation plausibly compensates for coverage changes after pruning while avoiding duplication of all attributes.
- Controlled evidence for selective sharing. For three levels, Table 7 reduces resident memory from 0.86 to 0.49 GB relative to independent LoD, with only a 0.09 dB PSNR decrease. It recovers much of the quality lost with fully shared attributes.
- Benefits persist without quantization. The uncompressed SAS-LoD rows in Table 10 outperform the reported LODGE implementation in Tables 8–9 on throughput, memory, and reconstruction metrics across all six scenes. Resident and working memory are reported separately.
Critical weaknesses
- The generality of the central finding is limited. The attribute analysis uses a specific hierarchy constructed by pruning and fine-tuning an optimized representation. Preserving positions, rotations, and appearance while repairing coverage is intuitive under this setup. It remains unclear whether four adaptation parameters remain adequate under stronger simplification or different geometry and appearance. The evidence supports this configuration more strongly than a general principle of Gaussian LoD.
- Large-scale application benefits are not fully demonstrated. The outdoor evaluation uses one chunk from each of only two scenes, with out-of-core streaming disabled. It does not measure complete-scene loading, caching, navigation, or peak memory. Uncompressed SAS improves outdoor throughput over LODGE by approximately 5–8%, while both already run at high frame rates; total memory falls by approximately 31–35%. Whether this enables materially larger scenes or fewer loading stalls remains unestablished.
- Comparable compression baselines are incomplete. Compression contributes substantially to resident-memory reduction in Table 1. Figure 7(d) does include quantized independent-LoD memory, so this control is not entirely absent. However, the paper lacks a complete quality–throughput–memory comparison with competing LoD methods under comparable compression and on-demand decoding. Such results would isolate the representation’s advantage more convincingly.
Questions for authors
Q1. Coverage of large-scale scenes.
Why is the outdoor evaluation restricted to a single chunk from each of Campus and SmallCity, and how were these chunks selected? Could the authors evaluate all chunks under the same protocol and report both per-chunk results and scene-level aggregates for rendering quality, throughput, and memory? This would help establish whether the reported benefits generalize across the complete scenes. Please distinguish evaluating all chunks individually from rendering the complete scene with streaming enabled.
Q2. Practical applications and significance.
Which practical applications would materially benefit from the proposed quality–speed–memory trade-off? In the outdoor experiments, uncompressed SAS-LoD improves throughput over LODGE by approximately 5–8% and reduces total memory by approximately 31–35%, while both methods already achieve high frame rates. Could the authors demonstrate a concrete setting where these savings enable a previously infeasible workload—for example, rendering a larger scene within a fixed GPU-memory budget—or substantially improve deployment performance? Such evidence would clarify the practical significance of the method beyond relative metric improvements.
Q3. Comparable compression baselines.
A substantial portion of the resident-memory savings comes from compression. Have competing LoD representations been evaluated with comparable quantization and selective decoding? Figure 7(d) provides a quantized independent-LoD memory comparison, but could the authors additionally report rendering quality, throughput, and resident/working memory under matched compression settings? This would help separate the benefits of selective attribute sharing from those of the compression pipeline.
Q4. Novelty and contribution attribution.
Is the main contribution of SAS-LoD the empirical finding about which attributes require adaptation, the compact scale/opacity residual representation, or the selective decoding strategy? The existing attribute ablations are useful; could the authors clarify what is technically new and which controlled comparisons isolate its benefit from the underlying hierarchy construction and compression choices?
9 comments
Response to Reviewer ALsr (2/5)
C2. Single-chunk evaluation
We kindly clarify that evaluating a single chunk for each outdoor scene adheres to the standard evaluation protocol of the Hierarchical-3DGS dataset. These chunks are the official evaluation benchmarks designated by the original dataset authors and were not selected by us. Subsequent large-scale 3DGS approaches, including Octree-GS, FLoD, and LODGE, have also adopted this evaluation setup. We therefore follow this protocol to ensure a fair and reproducible comparison with prior work.
C3. Full-chunk evaluation (Large-scale scenes)
We appreciate this insightful suggestion. To verify that our performance gains generalize across complete large-scale scenes, we evaluate all individual chunks of Campus (21 chunks, 22.0K images) and Small City (4 chunks, 5.8K images) under the same per-chunk evaluation protocol. The aggregated results show that the benefits of SAS-LoD are not limited to single benchmark chunks, but consistently generalize across all chunks of large-scale scenes.
We use every 50th image for testing, and all methods are trained and evaluated using this split. We report scene-wise aggregated results in Tables R15 & R16 while providing per-chunk results in the revised manuscript (Sec. C.8) due to space limitations. Quality and FPS are averaged over all test views of all chunks, while memory is reported as the average and maximum over chunks.
Note that this experiment evaluates every chunk individually. Rendering the complete scene, both in-core and with chunk streaming, is evaluated separately in C4.
1. Comparison with baselines across all chunks of each scene
The scene-level aggregate results in Tables R15 & R16 demonstrate that SAS-LoD achieves the highest average rendering throughput and the lowest memory footprint across all chunks, while maintaining rendering quality comparable to the best-performing baseline (LODGE). These aggregated results confirm that the benefits of our method extend across complete large-scale scenes, consistent with our single-chunk evaluations.
Table R15. Quantitative results on all chunks of Campus scene.
| Method | Mem (GB, Avg / Max) | Quality | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| #G | FPS ↑ | Res ↓ | Work ↓ | Total ↓ | PSNR ↑ | SSIM ↑ | LPIPS ↓ | |||
| 3DGS | 1055K | 119 | 0.65 / 1.03 | 0.68 / 0.96 | 1.34 / 1.99 | 24.77 | 0.754 | 0.378 | ||
| H3DGS | 589K | 126 | 1.38 / 2.18 | 0.51 / 0.66 | 1.89 / 2.83 | 24.72 | 0.750 | 0.384 | ||
| Octree-GS | 288K | 111 | 0.55 / 0.78 | 0.53 / 1.02 | 1.08 / 1.58 | 24.60 | 0.737 | 0.407 | ||
| FLoD | 230K | 142 | 1.00 / 1.28 | 0.40 / 0.47 | 1.40 / 1.67 | 24.83 | 0.731 | 0.416 | ||
| CLOD-3DGS | 394K | 209 | 0.65 / 1.03 | 0.35 / 0.46 | 1.00 / 1.49 | 24.49 | 0.747 | 0.390 | ||
| LODGE | 407K | 175 | 1.08 / 1.70 | 0.46 / 0.63 | 1.54 / 2.27 | 25.12 | 0.758 | 0.372 | ||
| CLoD-GS | 424K | 131 | 0.49 / 0.60 | 0.65 / 0.81 | 1.14 / 1.32 | 25.09 | 0.754 | 0.383 | ||
| Ours w/o quant. | 328K | 208 | 0.68 / 1.07 | 0.33 / 0.43 | 1.01 / 1.46 | 25.15 | 0.758 | 0.372 | ||
| Ours | 328K | 210 | 0.08 / 0.12 | 0.33 / 0.43 | 0.41 / 0.54 | 25.09 | 0.756 | 0.376 | ||
| Ours (single chunk) | 331K | 186 | 0.06 | 0.36 | 0.42 | 25.94 | 0.817 | 0.327 |
Table R16. Quantitative results on all chunks of Small City.
| Method | Mem (GB, Avg / Max) | Quality | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| #G | FPS ↑ | Res ↓ | Work ↓ | Total ↓ | PSNR ↑ | SSIM ↑ | LPIPS ↓ | |||
| 3DGS | 1026K | 105 | 0.76 / 0.78 | 0.59 / 0.64 | 1.35 / 1.42 | 24.55 | 0.768 | 0.302 | ||
| H3DGS | 409K | 149 | 1.53 / 1.57 | 0.32 / 0.33 | 1.85 / 1.89 | 24.40 | 0.761 | 0.315 | ||
| Octree-GS | 298K | 154 | 0.69 / 0.90 | 0.43 / 0.64 | 1.12 / 1.54 | 24.42 | 0.763 | 0.309 | ||
| FLoD | 269K | 197 | 0.99 / 1.04 | 0.27 / 0.28 | 1.26 / 1.32 | 24.29 | 0.745 | 0.344 | ||
| CLOD-3DGS | 418K | 193 | 0.76 / 0.78 | 0.26 / 0.27 | 1.02 / 1.05 | 24.40 | 0.767 | 0.306 | ||
| LODGE | 291K | 285 | 1.14 / 1.18 | 0.27 / 0.29 | 1.41 / 1.47 | 24.80 | 0.772 | 0.300 | ||
| CLoD-GS | 400K | 180 | 0.54 / 0.57 | 0.37 / 0.39 | 0.91 / 0.96 | 24.69 | 0.765 | 0.310 | ||
| Ours w/o quant. | 253K | 335 | 0.78 / 0.80 | 0.19 / 0.21 | 0.97 / 1.00 | 24.81 | 0.772 | 0.298 | ||
| Ours | 253K | 341 | 0.09 / 0.09 | 0.19 / 0.21 | 0.28 / 0.30 | 24.79 | 0.770 | 0.302 | ||
| Ours (single chunk) | 254K | 327 | 0.05 | 0.20 | 0.26 | 26.97 | 0.812 | 0.278 |
(Continued in the next comment)
Response to Reviewer ALsr (3/5)
C3. Full-chunk evaluation (Large-scale scenes) (continued)
2. Ablation study on LoD design across all chunks of each scene
Moreover, Table R17 presents an ablation study on LoD design conducted across all chunks of Campus and Small City. The results demonstrate that the effectiveness of our selective attribute sharing design also generalizes beyond a single benchmark chunk to complete scenes. Consistent with the single-chunk evaluation, SAS-LoD achieves rendering quality comparable to independent LoD while requiring about 70% of its resident memory. Compared with shared LoD, SAS-LoD achieves superior rendering quality at a similar resident memory cost.
Table R17. Ablation study on LoD designs across all chunks of Campus and Small City.
| Method | Mem (GB, Avg / Max) | Quality | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| #G | FPS ↑ | Res ↓ | Work ↓ | Total ↓ | PSNR ↑ | SSIM ↑ | LPIPS ↓ | |||
| Campus | ||||||||||
| Independent LoD | 341K | 198 | 1.02 / 1.61 | 0.34 / 0.43 | 1.36 / 2.02 | 25.33 | 0.759 | 0.371 | ||
| Shared LoD | 339K | 203 | 0.65 / 1.03 | 0.34 / 0.43 | 0.99 / 1.43 | 24.98 | 0.755 | 0.375 | ||
| SAS-LoD | 328K | 208 | 0.68 / 1.07 | 0.33 / 0.43 | 1.01 / 1.46 | 25.15 | 0.758 | 0.372 | ||
| Small City | ||||||||||
| Independent LoD | 263K | 315 | 1.08 / 1.12 | 0.20 / 0.21 | 1.28 / 1.32 | 24.89 | 0.774 | 0.297 | ||
| Shared LoD | 262K | 327 | 0.76 / 0.78 | 0.20 / 0.21 | 0.96 / 0.99 | 24.68 | 0.769 | 0.301 | ||
| SAS-LoD | 253K | 335 | 0.78 / 0.80 | 0.19 / 0.21 | 0.97 / 1.00 | 24.81 | 0.772 | 0.298 |
Response to Reviewer ALsr (1/5)
We sincerely appreciate your insightful comments. We have tried to respond to all these valuable concerns as below.
C1. Generality of SAS-LoD
1. Selective attribute sharing for a different LoD
While our attribute analysis in the paper is conducted using a pruning-based hierarchy construction pipeline, we observe that the underlying insight can also extend to other hierarchy construction pipelines. To verify this, we apply our selective attribute sharing strategy to H3DGS, an independent LoD baseline, which merges Gaussians bottom-up into a tree.
As shown in Table R12, integrating selective attribute sharing into H3DGS achieves rendering quality comparable to its original independent representation (25.24 dB vs. 25.30 dB) while reducing resident memory by 22.1%. In contrast, sharing all attributes across levels in H3DGS results in noticeable quality degradation (25.09 dB). This result indicates that our key finding is not restricted to a specific hierarchy construction pipeline and can generalize to other LoD architectures.
Table R12. Ablation study on LoD design with H3DGS (Campus).
2. Stronger simplification
To investigate the effectiveness of SAS-LoD under stronger simplification, we evaluate rendering quality using higher masking thresholds ϵ for LoD construction in Table R13. The results indicate that adapting only scale and opacity remains adequate even under stronger simplification. As the masking threshold increases, all three LoD designs contain fewer Gaussians at coarser levels (l>0), yielding progressively simpler representations. Across all masking thresholds, SAS-LoD maintains rendering quality comparable to independent LoD. In contrast, shared LoD consistently exhibits a noticeable quality gap compared to independent LoD, demonstrating that sharing all attributes fails to compensate for stronger simplification.
Table R13. Ablation study on the masking threshold for LoD construction (Small City).
3. Different geometry and appearance
We would like to clarify that our main paper demonstrates the effectiveness of SAS-LoD across diverse geometries and appearances in both outdoor (Hierarchical-3DGS) and indoor (Zip-NeRF) datasets. The outdoor scenes capture large-scale structures, such as buildings, vehicles, and trees along streetscapes, while the indoor scenes encompass multi-room environments (e.g., kitchens, living rooms, and bedrooms) with complex layouts and a wide variety of daily objects.
Furthermore, to complement the outdoor-only ablation study in our main paper (Fig. 7 and Table 9 in the manuscript), we provide an additional ablation study on LoD design using the indoor Zip-NeRF dataset in Table R14. SAS-LoD achieves rendering quality comparable to independent LoD (26.52 dB vs. 26.58 dB), whereas shared LoD incurs a consistent quality degradation (26.37 dB). This consistent result across indoor and outdoor scenes indicates that adapting only scale and opacity serves as a general adaptation strategy across diverse geometries and appearances.
Table R14. Ablation study on LoD designs using the Zip-NeRF dataset.