AnnotateMate: Open-source tool to analyse, modify COCO format annotation

---

🦚 Features

• One script does it all
  • analyse: Analyse and generate stats & plot
  • draw: Draw bbox/segm from annotation file
  • split: Split annotation file
  • merge: Merge multiple annotations files
  • remove categories: Remove categories/corresponding annotations from annotation file
  • filter annotation: Filter annotations by bbox min/max height/width
• Supported annotation format: COCO

👾 This repo is constantly updated with the new features - check in regularly for updates!

🔧 Installation

Install the requirements

pip3 install -r requirements.txt

🍽️ Getting started

Run the main.py with different command line options:

main.py [-h] [--log_file LOG_FILE] [--split SPLIT] [--remove_categories REMOVE_CATEGORIES [REMOVE_CATEGORIES ...]]
               [--output_annotation OUTPUT_ANNOTATION] [--draw] [--merge] [--image_dir IMAGE_DIR]
               [--output_draw OUTPUT_DRAW] [--stat] [--plot_stat] [--output_plot OUTPUT_PLOT] [--filter_annotation]
               [--min_width MIN_WIDTH] [--min_height MIN_HEIGHT] [--max_width MAX_WIDTH] [--max_height MAX_HEIGHT]
               annotation_file [annotation_file ...]

COCO annotation helper.

positional arguments:
  annotation_file       Path to the COCO annotation file(s)

options:
  -h, --help            show this help message and exit
  --log_file LOG_FILE   Path to the output log file
  --split SPLIT         Number of parts to split the JSON file into
  --remove_categories REMOVE_CATEGORIES [REMOVE_CATEGORIES ...]
                        List of category IDs to remove
  --output_annotation OUTPUT_ANNOTATION
                        Directory to save output annotation file
  --draw                Draw bounding boxes or masks
  --merge               Merge multiple coco-json files
  --image_dir IMAGE_DIR
                        Directory to image directory
  --output_draw OUTPUT_DRAW
                        Directory to save bbox/mask draw images
  --stat                Analyse annotation stats
  --plot_stat           Plot category-wise annotation counts
  --output_plot OUTPUT_PLOT
                        File path to save stats plot image
  --filter_annotation   Filter annotation by bbox min/max height/width
  --min_width MIN_WIDTH
                        Minimum width threshold for bbox
  --min_height MIN_HEIGHT
                        Minimum height threshold for bbox
  --max_width MAX_WIDTH
                        Maximum width threshold for bbox
  --max_height MAX_HEIGHT
                        Maximum height threshold for bbox

🕵️ Analyse

Analyse and generate statisctics and plot from annotation file

python3 main.py \
    <path/to/coco-json> \
    --stat \
    --log_file <path/to/save/stats-log.log> \
    --plot_stat \
    --output_plot <path/to/save/plot-file.png>

🎨 Draw

Draw bbox/segm from annotation file

python3 main.py \
    <path/to/coco-json> \
    --draw \
    --image_dir <path/to/image/dir> \
    --output_draw <path/to/save/draw/bbox-segm/dir>

✂️ Split

Split annotation file into n-parts

python3 main.py \
    <path/to/coco-json> \
    --split 4 \
    --output_annotation <path/to/save/split/annotation-josn>

🍻 Merge

Merge multiple coco annotations files

python3 main.py \
    <path/to/coco-json-2> <path/to/coco-json-2> \
    --merge \
    --output_annotation <path/to/save/merge/annotation-josn>

🧹 Remove Categories

Remove categories/corresponding annotations from annotation file

python3 main.py \
    <path/to/coco-json> \
    --remove_categories 2 3 \
    --output_annotation <path/to/save/remove-cat/annotation-josn>

🔍 Filter Annotation

Filter annotations by bbox min/max height/width

python3 main.py \
    <path/to/coco-json> \
    --min_width 20 \
    --min_height 20 \
    --max_width 150 \
    --max_height 160 \
    --output_annotation <path/to/save/filter/annotation-josn>

🐸 Reference

If you use this repo and like it, use this to cite it:

@misc{annotatemate,
      title={AnnotateMate: Open-source tool to analyse, modify COCO format annotation},
      author={Neelanjan Bhowmik},
      year={2024},
      url={https://github.com/NeelBhowmik/AnnotateMate}
    }