This script allows to convert the Cityscapes Dataset to Mircosoft's CoCo Format. The code heavily relies on Facebook's Detection Repo and Cityscapes Scripts.
The converted annotations can be easily used for Mask-RCNN or other deep learning projects.
Download the Cityscapes Dataset and organize the files in the following structure. Create an empty annotations directory.
data/
└── cityscapes
├── annotations
├── gtFine
│ ├── test
│ ├── train
│ └── val
└── leftImg8bit
├── test
├── train
└── val
main.py
inspect_coco.py
README.md
requirements.txt
pip install -r requirements.txt
To run the conversion execute the following
python main.py --dataset cityscapes --datadir data/cityscapes --outdir data/cityscapes/annotations
In order to run the visualization of the CoCo dataset you may run
python inspect_coco.py --coco_dir data/cityscapes
