@misc{indiciaeadfc83003743,
  title = {CompAdapt: Adaptable Composite Motion Modeling for Physics-Consistent Text-to-Video Generation},
  author = {Haoran Qin and Renlong Wu and Tianyu Huang and Yukang Ding and Hui Li and Wangmeng Zuo},
  year = {2026},
  url = {https://arxiv.org/abs/2609.21455},
  note = {Source identifier: 2609.21455}
}