Convert bdd100k annotation to Coco or VOC format , also support driverable area
- Download bdd100k dataset and save at $bdd100k/
'bdd100k_drivable_maps.zip'
'bdd100k_images.zip'
'bdd100k_labels_release.zip'
'bdd100k_lane_labels_trainval.zip'
- Do batch_split_annotation
python batch_split_annotation.py
- The output yaml file will save at $bdd100k/bdd100k.yaml , which can be used in my project Mobilenet-YOLO-Pytorch
show_images.sh