face_mask_detector

Real-time (~14 fps) face mask detector which accurately identifies bounding boxes of faces and labels whether a mask is worn (correctly) or not worn using a Faster-RCNN (with FPN backbone) architecture. Achieves 85% [email protected] on the kaggle dataset: https://www.kaggle.com/wobotintelligence/face-mask-detection-dataset.

Getting Started

Python>=3.6
Based on the runtime environment, i.e. CPU, CUDA GPU, please install the following packages:
- Install Detectron2>=0.2, please refer to its official INSTALL.md
- Install PyTorch>=1.4, please refer to its official website INSTALL
Install additional packages
```
pip install -r requirements.txt
```

Create symlinks of the downloaded datasets to ./datasets

cd ./datasets
ln -sfn /path/to/download/dataset1 dataset1

Name		Name	Last commit message	Last commit date
Latest commit History 27 Commits
balloon_test		balloon_test
datasets		datasets
output		output
.ftpconfig		.ftpconfig
.gitattributes		.gitattributes
.gitignore		.gitignore
Brian_BaselineModel.ipynb		Brian_BaselineModel.ipynb
Brian_Video_Test.ipynb		Brian_Video_Test.ipynb
README.md		README.md
Salar_RetinaNet_Detectron2.ipynb		Salar_RetinaNet_Detectron2.ipynb
__init__.py		__init__.py
dataset_process.ipynb		dataset_process.ipynb
facemask_dataset.py		facemask_dataset.py
inference.py		inference.py
mAP.py		mAP.py
register_dataset.ipynb		register_dataset.ipynb
requirements.txt		requirements.txt
result.txt		result.txt
test_mAP.py		test_mAP.py
train.py		train.py