BDD100k toolkit

Convert bdd100k annotation to Coco or VOC format , also support driverable area

1. Download bdd100k dataset and save at $bdd100k/

'bdd100k_drivable_maps.zip'
'bdd100k_images.zip'
'bdd100k_labels_release.zip' 
'bdd100k_lane_labels_trainval.zip'

2. Do batch_split_annotation

python batch_split_annotation.py

3. The output yaml file will save at $bdd100k/bdd100k.yaml , which can be used in my project Mobilenet-YOLO-Pytorch

show_images.sh