Research map / Video / Video generation
Infinite-Homography as Robust Conditioning for Camera-Controlled Video Generation
Min-Jung Kim, Jeongho Kim, Hoiyeong Jin, Junha Hyung, Choo Jaegul
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 |
|---|---|---|---|
| add infinite-homography warping module (ablation) | InfCam (with trajectory+intrinsic aug) · AugMCV test (different intrinsics) | PSNR: 21.866 -> 24.311 | Improved |
| add infinite-homography warping module (ablation) | InfCam (with trajectory+intrinsic aug) · AugMCV test (shared intrinsics) | LPIPS: 0.250 -> 0.198 | Improved |
| add trajectory augmentation (reverse-and-concatenate) (ablation) | InfCam baseline · AugMCV test (shared intrinsics) | PSNR: 19.228 -> 20.820 | Improved |
| add trajectory augmentation alone (ablation) | InfCam baseline · AugMCV test (different intrinsics) | PSNR: 18.480 -> 18.865 | No clear effect |
| add focal-length (intrinsic) augmentation (ablation) | InfCam with trajectory aug · AugMCV test (different intrinsics) | PSNR: 18.865 -> 21.866 | Improved |
| train on AugMCV instead of MCV+SCV mixture (ablation) | InfCam · WebVid | TransErr: 1.441 -> 0.839 | Improved |
| train on AugMCV instead of MCV+SCV mixture (ablation) | InfCam · WebVid | FVD: 235.958 -> 235.563 | No clear effect |
| use InfCam instead of TrajectoryCrafter | camera-controlled video generation · WebVid | RotErr: 5.007 -> 3.162 | Improved |
| use InfCam instead of TrajectoryCrafter | camera-controlled video generation · WebVid FF-Sync | TransErr: 0.502 -> 0.510 | No clear effect |
| use InfCam instead of ReCamMaster with interpolation | camera-controlled video generation · WebVid FF-Async | RotErr: 8.360 -> 2.718 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- ReCamMaster — ReCamMaster: Camera-Controlled Generative Rendering from A Single Video (ICCV 2025) · better on 3 of 3 shared results
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
- Beyond Inpainting: Unleash 3D Understanding for Stable Camera-Controlled Video Re-rendering (ECCV 2026 · similarity 0.89)
- AC3D: Analyzing and Improving 3D Camera Control in Video Diffusion Transformers (CVPR 2025 · similarity 0.88)
- SynCamMaster: Synchronizing Multi-Camera Video Generation from Diverse Viewpoints (ICLR 2025 · similarity 0.88)
- Training-free Camera Control for Video Generation (ICLR 2025 · similarity 0.87)
- NAMVIS: Next-Scale Autoregressive Multi-View Image Synthesis (NeurIPS 2026 · similarity 0.87)
- HumanVid: Demystifying Training Data for Camera-controllable Human Image Animation (NeurIPS 2024 · similarity 0.87)
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.
