Here we provide the code used for the accompanying publication: "The Stanford Drone Dataset is More Complex Than We Think: An Analysis of Key Characteristics"

We provide the code for preprocessing and analyzing the inD Dataset (in ./ind/), the Stanford Drone Dataset (in ./sdd/), and to run YNet with both datasets, all in line with our published work. 

In order to keep the repository at a manageable size and respect the requirement of requesting access for the inD Dataset (at the time of writing, at least), we omit the raw data for the datasets but provide the processing scripts and describe how to structure the annotation files within the directory in order for the scripts to work.

Similarly, we provide the code and some pretrained networks from our YNet experiments. While we omit PECNet, the data used was the same as for YNet, and we list the link to the author's code below. 

Further instructions can be found in the subdirectories README files, with explanations given in comments throughout the code. 
If you have any issues or questions please contact me at joshua.andle@maine.edu

Links
SDD:  https://cvgl.stanford.edu/projects/uav_data/
inD Dataset: https://www.ind-dataset.com/
PECNet and YNet: https://github.com/HarshayuGirase/Human-Path-Prediction