Trajectory-Prediction-Tools

This pytp package is a tool package for the Trajectory Prediction task. It contains tools for data augmentation, normalization, preprocessing, evaluation and so on.

Maintainer - Hao Xue

Installation

pip install pytp

Usage

Evaluation (calculate ADE and FDE)

from pytp.utils.evaluate import get_ade, get_fde

ade = get_ade(predicted, ground_truth)
fde = get_fde(predicted, ground_truth)

Augmentation

from pytp.utils.augment import Augmentation

aug = Augmentation()

Requirements

Python >= 3.6

Citation

If you find this package useful, please consider citing:

@inproceedings{xue2019pedestrian,
  title={Pedestrian Trajectory Prediction Using a Social Pyramid},
  author={Xue, Hao and Huynh, Du Q and Reynolds, Mark},
  booktitle={Pacific Rim International Conference on Artificial Intelligence},
  pages={439--453},
  year={2019},
  organization={Springer}
}
@inproceedings{xue2019location,
  title={Location-Velocity Attention for Pedestrian Trajectory Prediction},
  author={Xue, Hao and Huynh, Du and Reynolds, Mark},
  booktitle={2019 IEEE Winter Conference on Applications of Computer Vision (WACV)},
  pages={2038--2047},
  year={2019},
  organization={IEEE}
}

To Do