Segementation Mask to COCO Json Converter

mask	original	json_visualization

1. Install dependencies

pip install -r requirements.txt

2. Input Folder Structure

mask
  |_ image_001.png
  |_ image_002.png
  |_ ...

3. Convert Segmentation to COCO Json annotation

python main.py -i mask

option:
-i : mask directory

4. Visualize COCO json annotation

To cross check generated segmentation data, visualize it by drawing on original image.

python visualize_segmentation.py -i image -a mask.json

option:
-i : original image directory
-a : json annotation file

5. Change class name and color

Modify config/config.yaml.

Reference

• https://github.com/chrise96/image-to-coco-json-converter