Research map / Video / Video generation
ObjCtrl-2.5D: Training-free Object Control with Camera Poses
Zhouxia Wang, Yushi Lan, Shangchen Zhou, Chen Change Loy
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 scale-wise mask dilation (ablation) | ObjCtrl-2.5D · ObjCtrl-Test | ObjMC: 124.37 -> 120.37 | Improved |
| add Shared Warping Latent (ablation) | ObjCtrl-2.5D · ObjCtrl-Test | ObjMC: 122.07 -> 120.37 | Improved |
| replace FreeTraj copy-pasting shared latent with SWL (ablation) | ObjCtrl-2.5D · ObjCtrl-Test | ObjMC: 138.22 -> 120.37 | Improved |
| replace copy-pasting with SWL (ablation) | ObjCtrl-2.5D · ObjCtrl-Test | FID: 248.39 -> 247.48 | No clear effect |
| add Layer Control Module (ablation) | CameraCtrl-SVD · qualitative | object motion control | Improved |
| use 2D trajectory with depth from conditional image instead of real 3D trajectory (ablation) | ObjCtrl-2.5D · DAVIS | ObjMC: 92.08 -> 91.42 | No clear effect |
| replace training-free PEEKABOO* with ObjCtrl-2.5D | PEEKABOO* · DAVIS | ObjMC: 128.05 -> 91.42 | Improved |
| replace training-free FreeTraj* with ObjCtrl-2.5D | FreeTraj* · ObjCtrl-Test | ObjMC: 158.39 -> 120.37 | Improved |
| replace training-based DragNUWA with ObjCtrl-2.5D | DragNUWA · DAVIS | ObjMC: 37.57 -> 91.42 | Hurt |
| replace training-based DragAnything with ObjCtrl-2.5D | DragAnything · ObjCtrl-Test | Dynamic Degree: 0.7436 -> 0.3461 | Hurt |
| replace PEEKABOO* with ObjCtrl-2.5D | PEEKABOO* · DAVIS | Dynamic Degree: 0.2472 -> 0.3483 | Improved |
Later papers that compared with this method
1 later accepted paper reports ObjCtrl-2.5D as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Zo3T: Zero-Shot 3D-Aware Trajectory-Guided Image-to-Video Generation via Test-Time Training (AAAI 2026 · better on 5 of 5 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 2 on most shared results.
- FreeTraj* — FreeTraj: Tuning-Free Trajectory Control via Noise Guided Video Diffusion (IJCV 2026) · better on 6 of 6 shared results
- DragAnything — Drag Anything: Motion Control for Anything using Entity Representation (ECCV 2024) · better on 1 of 6 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
- PoseTraj: Pose-Aware Trajectory Control in Video Diffusion (CVPR 2025 · similarity 0.87)
- I2VControl-Camera: Precise Video Camera Control with Adjustable Motion Strength (ICLR 2025 · similarity 0.87)
- MotionPro: A Precise Motion Controller for Image-to-Video Generation (CVPR 2025 · similarity 0.86)
- 3DTrajMaster: Mastering 3D Trajectory for Multi-Entity Motion in Video Generation (ICLR 2025 · similarity 0.86)
- LeviTor: 3D Trajectory Oriented Image-to-Video Synthesis (CVPR 2025 · similarity 0.86)
- SG-I2V: Self-Guided Trajectory Control in Image-to-Video Generation (ICLR 2025 · similarity 0.86)
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.
