Bothered by the fact that it is generally difficult to find a parking slot around where I live, I hence had the attempt of detecting available parking space using some deep learning models. Decided to give it a real quick shot after seeing this post.
conda create -n MaskRCNN python=3.6
pip install -r requirement.txt
pip install git+https://github.com/philferriere/cocoapi.git#subdirectory=PythonAPI
git clone https://github.com/matterport/Mask_RCNN.git
cd Mask_RCNN
python3 setup.py install
Download the pre-trained models from here;
and copy it into folder models.
Sadly, running jupyter notebook from within a conda environment does not imply your notebook also runs in the same environment. Thankfully, there’s an easy fix for that, namely nb_conda, and you’ll get it using:
conda install nb_conda
jupyter notebook
Open scripts/mask-rcnn-healthcheck.ipynb