LabelMe Instance Segmentation Visualizer

This tool visualizes instance segmentation masks from LabelMe JSON format. It supports overlaying the segmentation masks on original images and grouping instances by their group_id. The tool also supports parallel processing for improved performance.

Features

• Visualizes LabelMe instance segmentation masks with color-coded instances.
• Supports blending segmentation masks with original images if specified.
• Groups instances by group_id and outlines their contours in white.
• Displays instance labels within each segmented area.
• Utilizes multiprocessing for faster processing of large datasets.

Requirements

• Python 3.6 or higher
• Required Libraries:
  • numpy
  • opencv-python
  • tqdm

You can install dependencies using the following command:

pip install numpy opencv-python tqdm

Usage

Command Syntax

python visualize_labelme.py <input_dir> <output_dir> [--src <src_dir>] [--workers <num_workers>]

Arguments

• input_dir (required): Directory containing LabelMe JSON files.
• output_dir (required): Directory where the generated segmentation mask images will be saved.
• --src (optional): Directory containing original images for blending with segmentation masks.
• --workers (optional): Number of parallel workers to use (default is the number of available CPU cores).

Example Commands

Visualize segmentation masks only

python visualize_labelme.py input_dir output_dir

Visualize segmentation masks overlaid on original images

python visualize_labelme.py input_dir output_dir --src path/to/images --workers 4

Output

• The generated segmentation masks are saved in .png format with the same filename as the original .json file.
• Each instance is displayed in a unique color, grouped by group_id.
• Segmentation contours are displayed in white.
• Labels are displayed at the center of each segmented area.

Example Output

• Example Input JSON:

{
    "shapes": [
        {
            "label": "marking_arrow",
            "points": [
                [1471, 991], [1474, 991], [1474, 993], [1475, 993],
                [1475, 996], [1473, 996], [1473, 995], [1471, 995]
            ],
            "group_id": 54,
            "shape_type": "polygon",
            "flags": {}
        }
    ],
    "imagePath": "example_image.jpg",
    "imageHeight": 1080,
    "imageWidth": 1920
}

Notes

• Ensure that imagePath in your JSON files corresponds correctly to the image files in your --src directory.
• For optimal performance, set --workers based on the number of available CPU cores.

License

This project is licensed under the MIT License.

Contact

If you have any questions or suggestions, please feel free to reach out!