motion dynamics
Motion dynamics studies the behavior of moving objects and their interactions, essential for developing AI systems that can understand and predict motion in real-world environments, such as in robotics or video analysis.
- Deep Compositional Phase Diffusion for Long Motion Sequence Generation
- Disentangled Concepts Speak Louder Than Words: Explainable Video Action Recognition
- Generative Pre-trained Autoregressive Diffusion Transformer
- Learning Temporal 3D Semantic Scene Completion via Optical Flow Guidance
- MotionRAG: Motion Retrieval-Augmented Image-to-Video Generation
- Non-stationary Equivariant Graph Neural Networks for Physical Dynamics Simulation
- RoboScape: Physics-informed Embodied World Model
- Track3R: Joint Point Map and Trajectory Prior for Spatiotemporal 3D Understanding
- Virtual Fitting Room: Generating Arbitrarily Long Videos of Virtual Try-On from a Single Image
- WorldWeaver: Generating Long-Horizon Video Worlds via Rich Perception