Mask R-CNN is a framework for object instance detection and segmentation.
This repository has a Mask R-CNN implementation written in Keras.
It contains a Jupyter Notebook for training a model, and a REST API for detecting objects.
Start the notebook environment to train your model by
cding into the
docker directory and running
You can access the Jupyter notebook server on http://localhost:8889
After you complete training, you can start the REST API by running
docker-compose -f docker-compose.yml -f docker-compose.api.yml up