Research map / Video / Video segmentation and tracking
OCK: Unsupervised Dynamic Video Prediction with Object-Centric Kinematics
Yeon-Ji Song, Jaein Kim, Suhyung Choi, Jin-Hwa Kim, Byoung-Tak 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 |
|---|---|---|---|
| add Object Kinematics (Joint-OCK) over SlotFormer | slot transformer · OBJ3D | PSNR: 33.083 -> 35.125 | Improved |
| add Object Kinematics (Cross-OCK) over SlotFormer | slot transformer · MOVi-C | PSNR: 19.482 -> 21.040 | Improved |
| add Object Kinematics (Cross-OCK) over SlotFormer | slot transformer · MOVi-D | PSNR: 20.675 -> 22.338 | Improved |
| add Object Kinematics (Joint-OCK) over SlotFormer | slot transformer · MOVi-E | PSNR: 21.269 -> 22.394 | Improved |
| add Object Kinematics (Cross-OCK) over SlotFormer | slot transformer · Waymo Open | PSNR: 19.127 -> 25.979 | Improved |
| add Object Kinematics (Cross-OCK) over SlotFormer | slot transformer · MOVi-C | FG-ARI: 0.240 -> 0.339 | Improved |
| add Object Kinematics (Joint-OCK) over SlotFormer | slot transformer · MOVi-B | PSNR: 21.329 -> 21.646 | Mixed |
| use Cross-OCK rather than Joint-OCK | OCK · MOVi-D | LPIPS: 0.282 -> 0.236 | Improved |
| use OCVP transformers instead of OCK on complex scenes | OCVP-Seq · MOVi-D/E | PSNR | Hurt |
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.
- SlotFormer — SlotFormer: Unsupervised Visual Dynamics Simulation with Object-Centric Models (ICLR 2023) · better on 8 of 8 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
- ObjectForesight: Predicting 3D Object Trajectories from Human Videos (ECCV 2026 · similarity 0.89)
- Understanding Object Dynamics for Interactive Image-to-Video Synthesis (CVPR 2021 · similarity 0.88)
- Structured Object-Aware Physics Prediction for Video Modeling and Planning (ICLR 2020 · similarity 0.88)
- Temporally Consistent Object-Centric Learning by Contrasting Slots (CVPR 2025 · similarity 0.87)
- VideoJAM: Joint Appearance-Motion Representations for Enhanced Motion Generation in Video Models (ICML 2025 · similarity 0.86)
- PlaySlot: Learning Inverse Latent Dynamics for Controllable Object-Centric Video Prediction and Planning (ICML 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.
