Mask-RCNN object detection

Disclosure

This code base was based from the tutorial below, and only briefly tested (do not expect it to run)

Tutorial

https://machinelearningmastery.com/how-to-train-an-object-detection-model-with-keras/

Mask-RCNN repo

https://github.com/matterport/Mask_RCNN

Kangaroo dataset

https://github.com/experiencor/kangaroo.git

Setup

Set up instructions (tested only on MacOS)

git clone https://github.com/matterport/Mask_RCNN.git
cd Mask_RCNN
python setup.py install
pip show mask-rcnn

If testing the kangaroo dataset, clone the following and put the kangaroo folder inside obj_detection:

git clone https://github.com/experiencor/kangaroo.git

And finally,

pip install keras
pip install scikit-image

Training the model

We use transfer learning to train the model; the base model we use is from the mask RCNN model trained on the COCO dataset. You can download the weights of the model from here, and place them inside the obj_detection folder.

To continue training the model,

python train.py

Evaluate the model

python evaluate.py