aashi7 · GitHub

@article{Manglik2019,
  archivePrefix = {arXiv},
  arxivId = {1903.09102},
  author = {Manglik, Aashi and Weng, Xinshuo and Ohn-bar, Eshed and Kitani, Kris},
  eprint = {1903.09102},
  journal = {arXiv:1903.09102},
  title = {{Forecasting Time-to-Collision from Monocular Video: Feasibility, Dataset and Challenges}},
  url = {https://arxiv.org/pdf/1903.09102.pdf},
  year = {2019}
}

We used Faster RCNN to get the bounding boxes of people. Please see preprocessing/getBoundingBoxes.

Read the original on github.com ↗