Computer vision project for personal learning.
The applications are located at apps directory and each one has it's own readme.md file.
Detects multiple people and verifies if each person is wearing a face mask or not in real time.
- Python >=3.6 (CPython), PIP
- OS: Windows, Linux, Mac
- Minimum CPU dual core @1.9GHz
Install the dependencies and download the datasets/models.
pip install -r requirements/requirements.txt
Note: if you get an error when installing the dlib package, follow this stackoverflow answer.
- Download the datasets/models from pyimagesearch.
- Change the directory and filenames by the
conf/settings.pyfile.
python main.py --settings=conf.settings
You can override the conf/settings.py file by a custom one in conf/env (it will by ignored by git).
python main.py --settings=conf.env.dev
Usage of www.pyimagesearch.com scripts.
python train_mask_detector.py --dataset data/input/dataset
| Term | Description |
|---|---|
| Computer vision | Science dicipline that make computers understand images of the real world. |
| SSD | Single shot detector. |
| Bounding box | Subarea in a image that may contain a detected object. |
| ROI | Region of interest. See Bounding box. |
| CNN | Convultional Neuronal Network. It's the preferred kind of network for computer vision. |
| Centroid | Center point of a geometry object. |
| Mobilenet | An effitient CNN for mobile vision applications. |
Code styling is mandatory. The standards are the following:
- Clean Code by Robert C. Martin
- Pocoo styleguide
- PEP 20
- English
- The maximum line length is 120 characters, but you must try to decrease it.
- Use type hints whenever is possible.
Follow the .pylintrc file
The models, dataset and algorithms are from www.pyimagesearch.com tutorials.