Dear visitor,

If you think about using this software - there are better alternatives out there that do the same (and much much more) and are actively maintained. I recommend you to check out fiftyone:

• https://voxel51.com/docs/fiftyone/

What is this

This tool given a COCO annotations file and COCO predictions file will let you explore your dataset, visualize results and calculate important metrics.

Running the explorer on example data

You can use the predictions I prepared and explore the results on the COCO validation dataset. The predictions are coming from a Mask R-CNN model trained with mmdetection.

1. Download (and extract in project directory) the labels, annotations and images:

https://drive.google.com/open?id=1wxIagenNdCt_qphEe8gZYK7H2_to9QXl

2. Setup using docker

sudo docker run -p 8501:8501 -it -v "$PWD"/coco_data:/coco_data i008/cocoexp:latest  \
    --coco_train /coco_data/ground_truth_annotations.json \
    --coco_predictions /coco_data/predictions.json  \
    --images_path /coco_data/images/

2. Setup using conda

conda env update
conda activate cocoexplorer
streamlit run coco_explorer.py -- --coco_train ./coco_data/ground_truth_annotations.json --coco_predictions ./coco_data/predictions.json  --images_path ./coco_data/val2017/

2. Setup using pip

python3 -m venv .venv
. .venv/bin/activate
pip install -r requirements.txt
streamlit run coco_explorer.py -- --coco_train ./coco_data/ground_truth_annotations.json --coco_predictions ./coco_data/predictions.json  --images_path ./coco_data/val2017/

3. go to http://localhost:8501

Running on your own data

In the same way you can explore your own results. Just follow the official COCO dataset format for annotations and predictions.

Examples