- Rating
- 2
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 2
Read review
Paper summary
FVVSplat aims to reconstruct dynamic Gaussian scenes from uncalibrated multi-view video streams using adaptive position sampling, bidirectional deformation, and linear attention. The reported quality and speed improvements are promising, and the problem has clear practical value. However, the distinction from StreamSplat is insufficiently established, several essential parts of the multi-view pipeline are unclear, and the experiments do not fully support the claims about streaming stability and efficiency.
Critical strengths
-
The paper addresses a useful problem. Reducing camera calibration requirements and per-frame optimization would make free-viewpoint video reconstruction easier to deploy.
-
The reported results are strong across three benchmarks. On N3DV, the method reaches 34.82 dB PSNR with a reported processing time of 0.26 seconds per frame.
-
The component ablations are useful. Table 3 examines all three proposed designs, and the visual comparisons show improvements in several challenging regions.
Critical weaknesses
-
The novelty needs stronger justification. StreamSplat already includes probabilistic position prediction, bidirectional deformation, and online Gaussian fusion. Equation 9 also follows established linear attention. The paper needs clearer attribution and direct feed-forward comparisons. The contribution list even retains the name “StreamSplat.”
-
The uncalibrated multi-view pipeline is underspecified. Section 4 mainly describes individual frames and temporal pairs. It does not clearly explain how different cameras are aligned, how their Gaussians share coordinates, or how target viewpoints are specified under the orthographic formulation.
-
The sampling bounds appear inconsistent with the depth mapping. The paper uses , but allows a sampling radius up to 1.2. The stated sampling interval can therefore cross , producing a singularity or negative depth. Any additional constraint needs to be specified.
-
The evaluation split is not sufficiently clear. Section 5.1 states that official splits are used, but does not specify input camera counts, held-out target views, or whether scenes overlap between training and testing. These details are essential for interpreting the large gains.
-
Long-term streaming stability is not demonstrated. Framewise image metrics and selected visual examples do not establish reduced temporal drift. The paper also lacks a clear pruning policy and measurements of Gaussian count or memory usage as the sequence grows.
-
The efficiency discussion overstates the evidence. In Table 4, efficient attention is slower at 480P, 720P, and 960P; it becomes faster only at 1080P. The timing scope is also unclear, including whether depth estimation, DINOv2 features, and rasterization are all counted.
Questions for authors
-
Which components are new relative to StreamSplat, and how does FVVSplat compare under matched conditions?
-
How are camera coordinates aligned and held-out target viewpoints rendered without input calibration?
-
What prevents sampled depth offsets from crossing the singularity at ?
-
Please specify the camera, scene, and temporal splits, including whether target images enter the model.
-
Can you provide long-sequence temporal results and a complete runtime and memory breakdown?