All data used for this project are available via the following link : data.
A data set has been created for each simulated environment.
A data set is organized as follows:
- folders
<data>/
that contain images related to "data
" (for example,vanishing-point/
contains images with annotated vanishing point); - a folder
json/
that contains all annotations (for example,json/vanishing-point.json
contains the annotations of all vanishing points).
JSON files contain relative links to images. Make sure to adapt these links if you want to train Deep Learning models on a GPU cluster, for example!