- Rating
- 3
- Confidence
- 2
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 2
Read review
Paper summary
WorldCast combines projected geometric guidance with geometry-aligned multi-view character memory to generate identity-consistent videos under large-angle camera motion. It introduces CLACT and demonstrates improved camera adherence and cross-view consistency.
Critical strengths
- CLACT provides diverse large-angle camera trajectories and multi-view character observations, supporting joint evaluation of camera control and character consistency.
- Quantitative results and ablations support the effectiveness of GAMA, whose removal increases MEt3R from 0.228 to 0.333 (Table 2).
Critical weaknesses
- How are real-world anchors acquired, and how does alignment handle collinear trajectories or anchors mapped to the same target frame?
- Can you report fixed-scale camera errors on synthetic scenes and character consistency against held-out views?
- How does GAMA compare with standard multi-reference attention trained on the same data?
Questions for authors
see weakness
5 comments
Response to Reviewer 2PoH (2/2): Camera Accuracy, Held-Out Views, and Controlled Comparisons
W2.1. Fixed-scale camera errors on synthetic scenes.
Following the reviewer's suggestion, we evaluate fixed-scale camera errors on the synthetic scenes in our evaluation benchmark. All methods use the same VGGT-Omega estimator and a common temporal grid.
For each case, we determine a reference scale from the ratio of rendered ground-truth character depth to the corresponding depth estimated by VGGT-Omega on the initial frame. This scale is shared across all methods and held fixed during trajectory alignment, where only rotation and translation are fitted. We additionally report Sim(3)-aligned RMSE, which allows global scale adjustment, and the mean scale ratio (ideal: 1). The GT-render reference follows the same estimation and evaluation protocol.
| Method | Fixed-scale RMSE (cm) ↓ | Sim(3) RMSE (cm) ↓ | Scale Ratio (ideal: 1) |
|---|---|---|---|
| GT-render reference | 12.25 | 3.14 | 0.996 |
| CamI2V | 120.95 | 112.12 | 1.307 |
| DualCamCtrl | 128.94 | 108.82 | 1.463 |
| GEN3C | 47.89 | 21.47 | 0.892 |
| Uni3C | 35.72 | 8.43 | 0.998 |
| Lyra 2.0 | 57.28 | 35.70 | 1.140 |
| ViewCrafter | 134.22 | 128.97 | 1.644 |
| ViewCrafter* | 139.36 | 133.15 | 1.995 |
| MVGenMaster | 52.05 | 13.73 | 1.035 |
| MVGenMaster* | 49.30 | 30.03 | 1.200 |
| ReCamMaster | 116.44 | 112.62 | 1.190 |
| DepthDirector | 114.12 | 97.21 | 2.308 |
| CameraAnything | 81.82 | 72.03 | 1.125 |
| WorldStereo | 59.66 | 42.90 | 1.175 |
| WorldCast | 18.54 | 6.09 | 1.069 |
WorldCast achieves the lowest position RMSE among generation methods under both fixed-scale (18.54 cm) and Sim(3) alignment (6.09 cm). These results confirm its camera-control advantage even when scale adjustment is not permitted during trajectory alignment.
W2.2. Character consistency against held-out views.
Following the reviewer's suggestion, we further evaluate character consistency against held-out views on synthetic scenes. We select four held-out GT views per case, each offset by approximately 45° from the conditioning anchor views. These held-out views are never used as inputs for video generation.
Using these held-out GT views as references, we evaluate character consistency with DINO, CLIP, ArcFace, and MEt3R.
| Method | DINO ↑ | CLIP ↑ | ArcFace ↑ | MEt3R ↓ |
|---|---|---|---|---|
| CamI2V | 0.712 | 0.868 | 0.097 | 0.718 |
| DualCamCtrl | 0.637 | 0.854 | 0.070 | 0.671 |
| GEN3C | 0.806 | 0.927 | 0.302 | 0.577 |
| Uni3C | 0.796 | 0.929 | 0.262 | 0.593 |
| Lyra 2.0 | 0.772 | 0.921 | 0.195 | 0.616 |
| ViewCrafter | 0.636 | 0.843 | 0.112 | 0.759 |
| ViewCrafter* | 0.579 | 0.820 | 0.091 | 0.779 |
| MVGenMaster | 0.659 | 0.859 | 0.152 | 0.634 |
| MVGenMaster* | 0.691 | 0.881 | 0.173 | 0.581 |
| ReCamMaster | 0.753 | 0.916 | 0.235 | 0.621 |
| DepthDirector | 0.767 | 0.902 | 0.297 | 0.643 |
| CameraAnything | 0.731 | 0.918 | 0.224 | 0.608 |
| WorldStereo | 0.729 | 0.908 | 0.176 | 0.645 |
| WorldCast | 0.910 | 0.962 | 0.419 | 0.461 |
WorldCast achieves the highest DINO, CLIP, and ArcFace similarities against held-out GT views and the lowest MEt3R error. These results demonstrate that WorldCast not only preserves consistency with the input character observations but also maintains character consistency at previously unseen viewpoints.
W3. Controlled comparison with standard multi-reference attention.
Following the reviewer's suggestion, we compared GAMA with standard multi-view memory attention using the same Wan2.1-I2V-14B backbone, training data, four character anchors, and training settings, including random anchor dropout.
| Method | ATE ↓ | RTE ↓ | RRE (°) ↓ | ArcFace ↑ | DINO ↑ | CLIP ↑ | MEt3R ↓ | Subj. ↑ | IQ ↑ |
|---|---|---|---|---|---|---|---|---|---|
| Standard Multi-View Memory Attention | 0.104 | 0.025 | 0.886 | 0.396 | 0.694 | 0.821 | 0.291 | 0.909 | 0.716 |
| WorldCast (GAMA) | 0.065 | 0.020 | 0.896 | 0.555 | 0.719 | 0.830 | 0.228 | 0.917 | 0.719 |
Under identical conditions, GAMA improves ArcFace by 0.159 and reduces MEt3R and ATE by 21.6% and 37.5%, respectively, while maintaining comparable RRE. These results further demonstrate the effectiveness of GAMA's geometry-aware memory retrieval in improving character consistency and camera-control accuracy under identical input and training conditions.
We sincerely hope that our detailed clarifications, supplementary evaluations, and manuscript revisions have fully addressed your concerns. Thank you again for your thoughtful and constructive comments. We welcome any further questions, comments, or discussion.
Follow-up to Reviewer 2PoH
Thank you again for your constructive feedback and valuable suggestions. As discussion deadline approaches, we would like to briefly follow up on our previous responses and manuscript revisions, including the additional experiments and clarifications addressing your concerns. We hope our responses and revisions have addressed your concerns and would welcome any further comments or feedback.
Thank you again for your consideration.
Response to Reviewer 2PoH (1/2): Anchor Acquisition and Geometric Alignment
We thank the reviewer for recognizing the value of CLACT and the effectiveness of GAMA. We provide additional implementation details and evaluation results below.
W1.1. Real-world anchor acquisition.
For the 25 real-world cases, we capture multi-view observations of the same character. One frame is randomly selected as the initial input, and complementary front, back, and side observations are selected as character anchors.
W1.2. Alignment for collinear trajectories.
Collinear camera centers leave the global rotation underconstrained when alignment relies on position correspondences alone. Our implementation checks the number of distinct target correspondences and their spatial spread. For degenerate configurations with an available subject-center estimate, we determine the alignment using the shared initial-frame orientation and subject-center geometry:
The subject center is obtained during benchmark preprocessing through least-squares intersection of camera viewing rays. For trajectories with parallel or collinear viewing rays, preprocessing uses an available non-degenerate orbit trajectory from the same static scene to estimate this center.
W1.3. Multiple anchors mapped to the same target frame.
Anchors sharing a target-frame correspondence are retained as independent observations. Their correspondences, together with other valid matches, are used to estimate one global Sim(3) transform. Applying this transform to each anchor's original camera pose preserves the relative geometric structure of the anchor configuration.
When repeated matches leave too few distinct target correspondences or insufficient spatial spread, alignment uses the shared initial-frame orientation and the available subject center, following the procedure in W1.2.
The resulting transform is applied to all anchor cameras, including those excluded from the fitting subset. Each anchor is independently encoded and augmented with a geometric embedding computed from its aligned camera pose. All anchor tokens are then concatenated into character memory, allowing video queries to retrieve complementary appearance and geometric information from the individual observations.