- Rating
- 3(+1)
- Confidence
- 4
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 2
Read review
Paper summary
This paper introduces GeoRoute, a plug-in framework for improving temporal geometric consistency in pretrained video diffusion models through selective 3D-prior injection. The key observation is that 3D–video latent compatibility varies along the denoising trajectory and across frequency bands, motivating adaptive decisions about both when and what geometric information should be injected. GeoRoute therefore combines a sample-specific Timestep Router with a Spectral Router and a decoupled training strategy to redirect the denoising trajectory using a text-generated 3D prior. The paper also introduces GeoVideoBench, a reference-free benchmark measuring local rigidity, long-range 3D consistency, global shape stability, and topology consistency. Experiments on Wan, HunyuanVideo, and CogVideoX show consistent improvements over the corresponding base models, while comparisons against GeoVideo and VideoGPA suggest smaller gains over existing geometry-aware approaches.
Critical strengths
-
The adaptive geometry-injection formulation is intuitive and reasonably well motivated. The paper provides empirical evidence that 3D–video compatibility is concentrated within an intermediate denoising window, varies across samples, and changes across frequency bands. This directly motivates the Timestep Router and Spectral Router instead of introducing them purely as architectural heuristics.
-
The method is evaluated across multiple video backbones and supported by extensive ablations. GeoRoute improves all three tested backbones, and the ablation study separately examines fixed versus adaptive routing, spectral routing, incorrect or blank priors, and persistent versus transient injection. These experiments provide useful evidence that the gains are associated with correctly matched 3D information and adaptive routing.
Critical weaknesses
-
The improvement over the strongest geometry-aware baseline is quite small. On the shared CogVideoX backbone, GeoRoute only improves the VBench average from 96.42 for VideoGPA to 96.69 and GeoVideoBench Overall from 65.43 to 65.78. Several individual geometric metrics also differ by less than one point. Given this small margin, stronger independent evaluation is needed to establish that the improvement over existing geometry-aware methods is practically meaningful. Moreover, all four GeoVideoBench components differ from VideoGPA by less than one point (RSPS +0.63, L3RC +0.81, GSSC +0.40, PTC +0.05), and results appear to come from a single random seed. Without these, it is not possible to tell whether the reported margins exceed sampling noise.
-
Performance on genuinely dynamic scenes is not sufficiently characterized. The paper reports aggregate geometric-consistency metrics and provides several dynamic qualitative examples, but it does not separately evaluate static scenes, camera-motion-dominated scenes, articulated subjects, and complex human–object interactions. This distinction is important because the injected TRELLIS prior is a prompt-level 3D representation, and the difficulty of maintaining geometry can differ substantially between rigid objects and genuinely dynamic scenes.
-
The evaluation and practical-cost story is incomplete. GeoVideoBench is introduced by the same paper and achieves moderate human alignment, so the small SOTA differences would benefit from stronger independent evaluation. Moreover, the human preference study only compares GeoRoute with the corresponding base backbones, not directly against GeoVideo or VideoGPA. GeoRoute also introduces an 8–20% inference overhead and requires online TRELLIS generation of a 3D prior for every prompt. The evidence for GeoVideoBench is also weaker than presented. Its overall human correlation (ρ = 0.275) is on par with GPT-4o (0.287) and SSIM-sim (0.269), so it is not shown to be better aligned with human judgment than generic baselines, and both the alignment study and the user study rely on eight participants. GeoVideoBench also scores only the largest foreground object (a SAM 2 mask on monocular relative depth), which matches the object-level TRELLIS prior GeoRoute injects; background and scene-level geometry are not evaluated. Together with the absence of VBench Dynamic Degree, the current metrics cannot distinguish improved geometry from reduced motion.
Questions for authors
-
Could the authors provide a quantitative breakdown for static versus genuinely dynamic scenes? In particular, I would like to see separate results for static objects/camera motion, articulated human or animal motion, and human–object or multi-object interactions. Since VideoGPA has shown that geometry-aware training can also improve dynamic scenes, it would be important to verify that GeoRoute’s gains are not dominated by rigid or relatively static cases.
-
Could the authors provide stronger independent evaluation of the small gains over VideoGPA and GeoVideo? Since GeoRoute improves the strongest CogVideoX-based baseline by only a small margin, I would find direct human preference comparisons against VideoGPA/GeoVideo particularly useful, not only against the original backbone. An additional external human-aligned evaluator, such as VideoReward (Improving Video Generation with Human Feedback). And addtional 3D consistent metrics such as 3DCS, Epipolar Error etc.
6 comments
jX9X — Response 2/4
Q2. Critical Weakness 2
Performance on genuinely dynamic scenes is not sufficiently characterized. The paper reports aggregate geometric-consistency metrics and provides several dynamic qualitative examples, but it does not separately evaluate static scenes, camera-motion-dominated scenes, articulated subjects, and complex human–object interactions. This distinction is important because the injected TRELLIS prior is a prompt-level 3D representation, and the difficulty of maintaining geometry can differ substantially between rigid objects and genuinely dynamic scenes.
GeoRoute improves VB and RC in all three categories, including articulated motion and human–object or multi-object interactions. We compared Wan and Wan + GeoRoute on the same examples in three OpenVid subsets of 100 each, selected with ChatGPT assistance. Static covers static objects or camera motion, Articulated covers human or animal motion, and Interaction covers human–object or multi-object interactions.
R4. Results by scene category.
| Method | Scene | SC | BC | MS | VB | RSPS | L3RC | GSSC | PTC | RC |
|---|---|---|---|---|---|---|---|---|---|---|
| Wan | Static | 94.17 | 96.29 | 98.64 | 96.37 | 44.37 | 76.87 | 87.38 | 51.20 | 66.19 |
| +GeoRoute | Static | 98.43 | 98.62 | 98.96 | 98.67 | 54.24 | 85.47 | 89.04 | 50.96 | 69.03 |
| Wan | Articulated | 93.05 | 95.21 | 98.46 | 95.57 | 42.80 | 73.89 | 85.34 | 50.72 | 64.72 |
| +GeoRoute | Articulated | 97.97 | 97.85 | 98.67 | 98.16 | 53.29 | 83.17 | 88.96 | 52.04 | 69.07 |
| Wan | Interaction | 91.02 | 92.58 | 96.44 | 93.35 | 38.76 | 69.35 | 81.17 | 45.62 | 59.90 |
| +GeoRoute | Interaction | 97.45 | 97.28 | 98.60 | 97.78 | 53.13 | 83.02 | 88.76 | 51.89 | 68.90 |
The interaction subset has the largest improvement in both aggregates. These results show that the gains extend to the tested dynamic scenes. The geometric scores follow the principal-foreground-object protocol described in Q3.
Q3. Critical Weakness 3
The evaluation and practical-cost story is incomplete. GeoVideoBench is introduced by the same paper and achieves moderate human alignment, so the small SOTA differences would benefit from stronger independent evaluation. Moreover, the human preference study only compares GeoRoute with the corresponding base backbones, not directly against GeoVideo or VideoGPA. GeoRoute also introduces an 8–20% inference overhead and requires online TRELLIS generation of a 3D prior for every prompt. The evidence for GeoVideoBench is also weaker than presented. Its overall human correlation (ρ = 0.275) is on par with GPT-4o (0.287) and SSIM-sim (0.269), so it is not shown to be better aligned with human judgment than generic baselines, and both the alignment study and the user study rely on eight participants. GeoVideoBench also scores only the largest foreground object (a SAM 2 mask on monocular relative depth), which matches the object-level TRELLIS prior GeoRoute injects; background and scene-level geometry are not evaluated. Together with the absence of VBench Dynamic Degree, the current metrics cannot distinguish improved geometry from reduced motion.
Independent evaluation and direct human comparisons.
The external comparisons in Q1 provide evidence beyond GeoVideoBench. Our human preference study compares GeoRoute with its base backbones; the added comparison with VideoGPA uses automated VideoReward preferences, not a new human study against VideoGPA or GeoVideo.
Online prior generation and inference cost.
Including online TRELLIS generation for each prompt, GeoRoute adds 9.32%–23.23% end-to-end inference time, mainly for prior generation, compatibility evaluation, and prior-feature processing. The original denoising category included 3D-to-2D projection, tensor concatenation, and sparse/dense point-cloud conversion; R5 reports these operations separately under routing/prior processing.
R5. End-to-end inference latency in seconds.
| Method | TRELLIS | Routing/prior | Denoising | Others | Total | Overhead |
|---|---|---|---|---|---|---|
| Wan2.1 | - | - | 203.91 | 6.08 | 209.99 | - |
| Wan2.1 + GeoRoute | 7.08 | 32.74 | 204.17 | 5.89 | 249.88 | 19.00% |
| CogVideoX | - | - | 211.32 | 20.78 | 232.10 | - |
| CogVideoX + GeoRoute | 9.04 | 43.33 | 212.96 | 20.69 | 286.02 | 23.23% |
| HunyuanVideo | - | - | 1174.81 | 38.34 | 1213.15 | - |
| HunyuanVideo + GeoRoute | 17.19 | 95.18 | 1175.22 | 38.58 | 1326.17 | 9.32% |
Denoising itself increases by only 0.26–1.64 seconds across the three backbones. A single injection still requires constructing the geometric representation and evaluating compatibility before transferring it, which explains the additional processing cost.
jX9X — Response 3/4
Q3 continued — benchmark evidence and motion.
Benchmark scope, human alignment, and the eight-rater study.
GeoVideoBench evaluates local rigidity, long-range 3D consistency, global shape stability, and silhouette topology of the principal foreground object; its validation combines controlled geometric tests with moderate agreement with human judgments.
For multi-object scenes, PTC can capture topology changes visible in the retained foreground silhouette, including changes in holes, skeleton branches, and thin structures. The correctness of interactions themselves, including semantics and physical dynamics, is outside this scope. The protocol also does not assess full background or scene-level 3D geometry. VBench Background Consistency separately measures temporal background consistency, and the external evaluations in R2–R3 provide additional evidence beyond GeoVideoBench.
The controlled validation in original Appendix C.4 uses 300 rendered videos with known geometry from six articulated assets and two camera paths. Corruption severity correlates with RC at −0.672 using oracle geometry and −0.708 through the RGB pipeline. Original Tables 7–9 test sensitivity to targeted geometric corruption and nuisance changes in lighting, appearance, and camera motion. Reliability weights are fixed across methods, and original Table 10 reports gains under both equal weighting and RC.
Within our human-alignment study, RC has Spearman ρ = 0.275 and pairwise agreement of 66.0%. The cited GPT-4o (0.287) and SSIM-sim (0.269) correlations provide context from different tasks and protocols; they do not establish a ranking under a common evaluation. To make agreement among our raters explicit, we additionally calculated Krippendorff's alpha for the blinded preference study.
R6. Krippendorff's alpha with 95% confidence intervals.
| Criterion | Alpha | 95% CI |
|---|---|---|
| Visual quality | 0.3842 | [0.3227,0.4410] |
| Geometric consistency | 0.3385 | [0.2157,0.4532] |
Both estimates indicate modest inter-rater agreement. The preference and metric-alignment blocks use the same eight participants; the 16,534 criterion-specific preference judgments include ties and repeated assessments. Alpha measures agreement among raters, while metric–human correlation measures alignment between metric scores and human judgments.
Motion presence and reduced-motion controls.
GeoRoute maintains Dynamic Degree on all three backbones and improves VB and RC in every baseline motion group, including the high-motion group. We first report VBench Dynamic Degree, which measures whether motion is present.
R7. VBench Dynamic Degree (%).
| Backbone | Base | +GeoRoute |
|---|---|---|
| Wan | 56.27 | 56.54 |
| CogVideoX | 66.37 | 66.52 |
| HunyuanVideo | 59.34 | 59.38 |
Dynamic Degree is slightly higher with GeoRoute on each backbone. To test different motion levels, we divided Wan's VBench outputs into three groups using the mean magnitude of the largest 5% of inter-frame pixel flows. Each paired GeoRoute output retains its Wan counterpart's group, so the comparisons use the same examples.
R8. Results by baseline motion magnitude.
| Method | Motion | SC | BC | MS | VB | RSPS | L3RC | GSSC | PTC | RC |
|---|---|---|---|---|---|---|---|---|---|---|
| Wan | Low | 93.94 | 96.52 | 98.39 | 96.28 | 44.71 | 77.12 | 87.05 | 51.44 | 66.30 |
| +GeoRoute | Low | 98.68 | 98.39 | 99.21 | 98.76 | 53.98 | 85.72 | 89.28 | 50.72 | 68.99 |
| Wan | Mid | 92.83 | 95.42 | 98.22 | 95.49 | 42.56 | 73.65 | 85.61 | 50.96 | 64.83 |
| +GeoRoute | Mid | 98.19 | 97.64 | 98.43 | 98.09 | 53.04 | 82.93 | 88.72 | 52.29 | 69.06 |
| Wan | High | 92.77 | 93.27 | 99.16 | 95.06 | 43.44 | 71.59 | 87.92 | 50.48 | 64.93 |
| +GeoRoute | High | 96.89 | 97.64 | 98.25 | 97.60 | 52.58 | 81.28 | 88.28 | 52.66 | 68.79 |
The high-motion group improves by 2.54 VB and 3.86 RC, and both aggregates also improve in the low- and medium-motion groups. Individual dimensions do not improve uniformly: high-motion MS decreases by 0.91 points and low-motion PTC by 0.72 points.
The reduced-motion controls explain why motion and geometric consistency should be reported together. We randomly sampled 800 videos generated on the VBench prompt set and evaluated each original alongside slowed-resampling and repeated-single-frame versions.
R9. Motion controls.
| Version | Dynamic Degree (DD, %) | RSPS | L3RC | GSSC | PTC | RC |
|---|---|---|---|---|---|---|
| Original | 53.13 | 53.20 | 83.31 | 88.76 | 51.89 | 68.95 |
| Slowed | 11.38 | 53.67 | 84.22 | 89.43 | 53.27 | 70.02 |
| Repeated frame | 0.00 | 57.21 | 85.93 | 90.07 | 56.78 | 72.53 |
Reduced motion can increase geometric-consistency scores; a repeated frame is geometrically stable but contains no motion. GeoRoute maintains Dynamic Degree (R7), and VideoReward provides complementary motion-quality and text-alignment evidence (R2). Dynamic Degree measures motion presence, not exact amplitude or action-specific correctness.
jX9X — Response 1/4
Thank you for the careful review and for recognizing the empirical basis of our two routers, the improvements across Wan, HunyuanVideo, and CogVideoX, and the controlled ablations on routing, prior matching, and injection duration. We have added multi-seed comparisons, independent evaluations, and quantitative breakdowns for dynamic scenes to address your questions.
Q1. Critical Weakness 1
Multiple seeds and significance.
GeoRoute achieves higher VB and RC than VideoGPA in all five additional matched generation runs, and the improvements in VB and Motion Smoothness remain significant after Holm correction. The original single-seed results did not quantify variability across runs. We compared both methods on the shared CogVideoX backbone using five new generation seeds, calculated the aggregates separately for each seed, and report sample standard deviations, paired 95% confidence intervals, and two-sided paired t-tests. Holm correction covers all nine scores.
R1. Five-seed comparison. Δ denotes GeoRoute minus VideoGPA.
The mean improvements are 0.682 in VB and 0.669 in RC, and every component metric has a higher mean. VB and Motion Smoothness have Holm-adjusted p values of 0.0404. For RC, the unadjusted confidence interval is positive, while the Holm-adjusted p value is 0.2428. These additional runs support a consistent observed advantage over VideoGPA, with stronger statistical evidence for VB than for RC.
Independent evaluation.
The added independent geometric metrics and VideoReward comparison favor GeoRoute over VideoGPA on the shared CogVideoX backbone. We used the requested 3DCS and Epipolar Error metrics and the external VideoReward evaluator to compare the two methods directly.
R2. External evaluation against VideoGPA on CogVideoX. Win rates are automated VideoReward preferences for GeoRoute over VideoGPA.
GeoRoute reduces both geometric errors and obtains VideoReward win rates above 50% in visual quality, motion quality, and text alignment. These results provide evidence of improved geometry and perceptual quality beyond GeoVideoBench. The new direct comparison is against VideoGPA; the comparison with GeoVideo is reported in the original main results. Our human preference study compares GeoRoute with its base backbones, so R2 adds an automated external preference assessment rather than a new human study against VideoGPA or GeoVideo.
We also evaluated Wan and Wan + GeoRoute with PDI-Bench as a separate geometry benchmark.
R3. PDI-Bench evaluation of Wan and Wan + GeoRoute.
All five reported PDI-Bench metrics improve with GeoRoute, including a PDIS reduction from 0.7263 to 0.5688. Together, these evaluations supplement the original benchmark with independent evidence on both the base-model improvement and the comparison with VideoGPA.