Research map / Video / Video generation
ReCamMaster: Camera-Controlled Generative Rendering from A Single Video
Jianhong Bai, Menghan Xia, Xiao Fu, Xintao Wang, Lianrui Mu, Jinwen Cao, Zuozhu Liu, Haoji Hu, Xiang Bai, Pengfei Wan, Di Zhang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use frame concatenation instead of channel concatenation for video conditioning (ablation) | ReCamMaster · WebVid 1000 videos x 10 trajectories (conditioning ablation) | FID: 74.09 -> 57.10 | Improved |
| use frame concatenation instead of channel concatenation for video conditioning (ablation) | ReCamMaster · WebVid 1000 videos x 10 trajectories (conditioning ablation) | Mat. Pix.(K): 521.10 -> 906.03 | Improved |
| use frame concatenation instead of view concatenation for video conditioning (ablation) | ReCamMaster · WebVid 1000 videos x 10 trajectories (conditioning ablation) | FID: 80.51 -> 57.10 | Improved |
| use frame concatenation instead of view concatenation for video conditioning (ablation) | ReCamMaster · WebVid 1000 videos x 10 trajectories (conditioning ablation) | FVD-V: 177.68 -> 90.38 | Improved |
| use frame concatenation instead of view concatenation for video conditioning (ablation) | ReCamMaster · WebVid 1000 videos x 10 trajectories (conditioning ablation) | CLIP-T: 34.66 -> 34.53 | No clear effect |
| add noise to conditional video latent (ablation) | ReCamMaster · WebVid 1000 videos x 10 trajectories (training strategy ablation) | FID: 66.67 -> 65.17 | Improved |
| fine-tune only 3D attention layers (ablation) | ReCamMaster · WebVid 1000 videos x 10 trajectories (training strategy ablation) | FID: 66.67 -> 59.47 | Improved |
| drop source latent with 20% probability for T2V and I2V training (ablation) | ReCamMaster · WebVid 1000 videos x 10 trajectories (training strategy ablation) | FID: 66.67 -> 62.39 | Improved |
| combine noise, 3D attention tuning and latent dropping (ablation) | ReCamMaster · WebVid 1000 videos x 10 trajectories (training strategy ablation) | FVD: 171.80 -> 122.74 | Improved |
| train on high-quality diverse data instead of toy data (ablation) | ReCamMaster · WebVid 1000 videos x 10 trajectories (training data ablation) | FID: 69.35 -> 57.10 | Improved |
| train on high-quality diverse data instead of toy data (ablation) | ReCamMaster · WebVid 1000 videos x 10 trajectories (training data ablation) | FVD-V: 89.58 -> 90.38 | No clear effect |
Later papers that compared with this method
22 later accepted papers report ReCamMaster* as a baseline in their results tables; 20 of them report a better number on most of the results they share with it.
- BulletTime: Decoupled Control of Time and Camera Pose for Video Generation (CVPR 2026 · better on 8 of 8 shared results)
- CameraAnything: Refilming Videos with Arbitrary Camera Control (ECCV 2026 · better on 7 of 8 shared results)
- Unified Camera Positional Encoding for Controlled Video Generation (CVPR 2026 · better on 6 of 7 shared results)
- ReDirector: Creating Any-Length Video Retakes with Rotary Camera Encoding (CVPR 2026 · better on 7 of 7 shared results)
- FaceCam: Portrait Video Camera Control via Scale-Aware Conditioning (CVPR 2026 · better on 7 of 7 shared results)
- Plenoptic Video Generation (CVPR 2026 · better on 7 of 7 shared results)
- CRePE: Curved Ray Expectation Positional Encoding for Unified-Camera-Controlled Video Generation (NeurIPS 2026 · better on 6 of 6 shared results)
- World-R1: Reinforcing 3D Constraints for Text-to-Video Generation (ICML 2026 · better on 4 of 5 shared results)
14 more not listed.
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- SynCamMaster: Synchronizing Multi-Camera Video Generation from Diverse Viewpoints (ICLR 2025 · similarity 0.91)
- VD3D: Taming Large Video Diffusion Transformers for 3D Camera Control (ICLR 2025 · similarity 0.89)
- ReCapture: Generative Video Camera Controls for User-Provided Videos using Masked Video Fine-Tuning (CVPR 2025 · similarity 0.88)
- Wan-Move: Motion-controllable Video Generation via Latent Trajectory Guidance (NeurIPS 2025 · similarity 0.88)
- AC3D: Analyzing and Improving 3D Camera Control in Video Diffusion Transformers (CVPR 2025 · similarity 0.88)
- Training-free Camera Control for Video Generation (ICLR 2025 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
