- Rating
- 3(+1)
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 3
Read review
Paper summary
The paper proposes DIGS, a 3D Gaussian Splatting approach guided by diffusion priors. DIGS consists of four major components: (1) Group-Block-Diagonal (GBD) approximates Fisher information using blocks containing neighboring Gaussians, preserving some inter-Gaussian correlations. (2) Reliable Information Gain (RIG) measures the reliability of the diffusion-generated pseudo ground truths using Depth Anything 3 and computes the gain of reliable Fisher information. (3) Greedy view selection strategy progressively selects the pseudo view with the highest RIG. (4) Uncertainty-aware optimization emphasizes pseudo supervision where rendering uncertainty and estimated reliability are high. The experiments show that DIGS consistently outperforms the baselines qualitatively and quantitatively. Ablation studies also verify the individual effectiveness of GBD and RIG.
Critical strengths
- The proposed DIGS demonstrates consistent improvements over the NerfBusters, WildExplore, DL3DV benchmarks.
- The supplementary material includes high-quality rendered videos with plausible 3D consistency.
- The proposed GBD and RIG are compelling and mathematically sound.
Critical weaknesses
- Missing DA3 details. The paper should explicitly explain how point-wise confidence is computed by DA3. Explaining the full pipeline of DA3 is not necessary. However, since the reliability estimation is a core component of the proposed DIGS, the paper should at least provide intuitive explanation about how the geometry consistency is measured and used to guide the reliability estimation in DA3. Equation 4 along does not contain much information.
- Critical implementation details missing. The paper misses some implementation details, such as the input views in Table 2 and the input view selection protocol in Table 3. These details are critical for the evaluation.
- Insufficient baseline coverage. Diffusion-guided 3DGS is often used under sparse-view scenarios to fill the unseen parts. I believe these baselines are highly related to the paper and should be included for comparison. Examples include Deceptive-3DGS [1], RI3D [2], GSFixer [3], and many more. More critically, FreeFix also uses Fisher information to guide the diffusion denoising process. The paper includes the discussion with FreeFix but does not include the comparison.
- Insufficient evidence of extrapolation. The method is motivated by boosting the 3D consistency of pseudo extrapolated views. However, the paper does not provide sufficient evidence to show that DIGS improves the quality of pseudo views and rendered views with large viewpoint extrapolation. The paper has to at least include an additional figure with examples of the training views, selected extrapolated pseudo views, and rendered views at the same camera poses.
Overall, I find the method technically sound and novel. However, some important implementation details are missing, and the evaluation lacks key comparisons and analyses. Therefore, I recommend weak reject in its current form. I am open to raise my score if the above concerns are well-addressed during the rebuttal period.
- [1] Liu et al., “Deceptive-NeRF/3DGS: Diffusion-Generated Pseudo-Observations for High-Quality Sparse-View Reconstruction”, ECCV 2024.
- [2] Paliwal et al. “RI3d: Few-shot gaussian splatting with repair and inpainting diffusion priors”, ICCV 2025.
- [3] Yin et al. “GSFixer: Improving 3D Gaussian Splatting with Reference-Guided Video Diffusion Priors”, ICML 2026.
Questions for authors
- How many input views are used in Table 2? Are you using the full training dataset? If so, is there a reason to compare feed-forward methods under the sparse-view setting, while comparing with the single-scene methods only with the full datasets? Can you conduct the same limited-input experiments in Table 2?
- How are the input views selected in Table 3?
7 comments
W3 Comparison with more related works.
We provide more quantitative comparisons with related works in the table below, following the evaluation setting of Table 2 in our manuscript. While Deceptive-3DGS, RI3D and GSFixer also use diffusion models to improve 3DGS, they mainly focus on the sparse-view setting, where the test views are typically in the same distribution as the training views. In contrast, we consider a more challenging setting, where the test views are from distinct trajectories and out of the distribution of the training views. Moreover, these methods do not jointly consider the informativeness or reliability of the generated pseudo views. While FreeFix uses a Fisher information-based confidence map to guide the denoising process of the diffusion model, it does not consider selecting the most effective pseudo views. In contrast, DIGS selects pseudo views that are both informative and reliable.
| Method | NerfBusters | WildExplore | DL3DV | ||||||
|---|---|---|---|---|---|---|---|---|---|
| PSNR ↑ | SSIM ↑ | LPIPS ↓ | PSNR ↑ | SSIM ↑ | LPIPS ↓ | PSNR ↑ | SSIM ↑ | LPIPS ↓ | |
| 3DGS | 14.51 | 0.417 | 0.454 | 15.81 | 0.487 | 0.434 | 16.19 | 0.537 | 0.363 |
| Deceptive-3DGS | 14.82 | 0.433 | 0.441 | 16.11 | 0.509 | 0.426 | 16.76 | 0.574 | 0.351 |
| RI3D | 15.27 | 0.454 | 0.437 | 16.20 | 0.515 | 0.418 | 17.09 | 0.580 | 0.344 |
| GSFixer | 15.93 | 0.471 | 0.441 | 16.78 | 0.522 | 0.428 | 17.46 | 0.593 | 0.337 |
| FreeFix | 16.01 | 0.482 | 0.424 | 16.91 | 0.529 | 0.407 | 17.53 | 0.584 | 0.315 |
| Difix3D+ | 15.78 | 0.468 | 0.433 | 16.56 | 0.514 | 0.423 | 17.33 | 0.588 | 0.324 |
| ExploreGS | 16.29 | 0.479 | 0.434 | 16.96 | 0.534 | 0.413 | 17.65 | 0.592 | 0.319 |
| Ours | 17.15 | 0.523 | 0.381 | 17.78 | 0.586 | 0.349 | 18.14 | 0.623 | 0.279 |
As shown in the above table, Deceptive-3DGS, RI3D, GSFixer and FreeFix improve the overall reconstruction quality compared to 3DGS. Our method consistently outperforms all baselines, validating the importance of informative and reliable view selection when incorporating diffusion priors.
W4. Evidence of view extrapolation. We provide more evidence of view extrapolation in Figure 11 on page 10 of the updated PDF, which shows the training views, the selected extrapolated pseudo views, and the rendered and diffusion-generated images at the same camera poses. The selected extrapolated pseudo views expand the coverage of the training views, and the diffusion-generated images complete scene boundaries and remove rendering artifacts, improving reconstruction quality at extrapolated test views. By selecting informative and reliable pseudo views, the diffusion-generated images generally remain consistent with the scene, even under large viewpoint changes.
Q1: Number of views and limited-input experiments in Table 2. In Table 2 of our manuscript, we use the full training views of each dataset by default, and evaluate the reconstruction quality on extrapolative test views from distinct trajectories. The numbers of views for each dataset are listed in the table below.
| Dataset | # Scenes | Avg. # Training Views | Avg. # Test Views |
|---|---|---|---|
| NerfBusters | 12 | 181 | 183 |
| WildExplore | 8 | 186 | 296 |
| DL3DV | 12 | 172 | 172 |
As explained in W2, we compare with feed-forward methods under the sparse-view setting in Table 3, since they are mostly trained with limited input views. As suggested by the reviewer, we further conduct the limited-input experiments of Table 2 on NerfBusters with 16, 32, and 64 training views in the table below.
| Method | 16 views | 32 views | 64 views | All views | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| PSNR ↑ | SSIM ↑ | LPIPS ↓ | PSNR ↑ | SSIM ↑ | LPIPS ↓ | PSNR ↑ | SSIM ↑ | LPIPS ↓ | PSNR ↑ | SSIM ↑ | LPIPS ↓ | |
| 3DGS | 13.51 | 0.339 | 0.516 | 13.58 | 0.344 | 0.501 | 13.98 | 0.374 | 0.469 | 14.51 | 0.417 | 0.454 |
| Difix3D+ | 13.82 | 0.346 | 0.505 | 14.27 | 0.385 | 0.491 | 14.73 | 0.406 | 0.442 | 15.78 | 0.468 | 0.433 |
| ExploreGS | 14.01 | 0.358 | 0.489 | 15.15 | 0.417 | 0.460 | 15.69 | 0.442 | 0.431 | 16.29 | 0.479 | 0.434 |
| AnySplat | 11.57 | 0.276 | 0.591 | 11.93 | 0.291 | 0.590 | 12.09 | 0.306 | 0.593 | 12.06 | 0.311 | 0.596 |
| Ours | 14.37 | 0.376 | 0.469 | 16.29 | 0.467 | 0.443 | 16.56 | 0.484 | 0.417 | 17.15 | 0.523 | 0.381 |
As shown in the above table, per-scene optimization methods consistently outperform AnySplat across different numbers of input views. The 16-view extrapolative setting is extremely challenging, since the training views only provide limited coverage of the scene, making it difficult to improve the reconstruction quality even with diffusion priors. As the number of input views increases, diffusion-enhanced methods improve more significantly, while AnySplat remains poor. DIGS outperforms all baselines under all limited-input settings. Notably, DIGS with only 32 input views achieves reconstruction quality comparable to ExploreGS with the full training views, further validating the importance of informativeness- and reliability-aware view selection.
Q2: View selection in Table 3. In Table 3, we uniformly sample the input views from the full training views, which is a common practice in the sparse-view evaluation setting. All methods use the same input views for a fair comparison.
We really appreciate the reviewer's valuable time and effort. We hope our responses address the reviewer's concerns, and we will incorporate these discussions and results into the revised version. We would be happy to address any further questions, and we would appreciate it if the reviewer could consider raising the score if the concerns are addressed.
We appreciate Reviewer 8Pia for their positive comments and constructive suggestions. We address the reviewer's concerns below.
W1 Details on the use of DA3. We provide an intuitive explanation of how the geometric consistency is measured with DA3 and how it is used.
We will add these details on DA3 and the geometric consistency in the revised version.
W2 Implementation details.
We clarify the experimental settings and the reasons below.
We will add these details to further clarify our experimental settings.
[1] Wu, Jay Zhangjie, et al. "Difix3d+: Improving 3d reconstructions with single-step diffusion models."
[2] Kim, Minsu, et al. "Exploregs: Explorable 3d scene reconstruction with virtual camera samplings and diffusion priors."
[3] Jiang, Lihan, et al. "Anysplat: Feed-forward 3d gaussian splatting from unconstrained views."