Real-Time Face Mask Detection is a computer vision project that leverages machine learning techniques to detect whether a person is wearing a face mask in real-time. This project uses popular deep learning frameworks like TensorFlow and OpenCV, making it a valuable resource for applications related to health and safety, such as monitoring mask compliance in public spaces. Transfer learning was applied by using a pre-trained model MobileNetV2 to achieve reliable results.
The project consists of the following key files:
-
augmentation.py: Given that the dataset is not large enough, more data was created using thealbumentationlibrary in this script. -
validate_images.py: After data was created, this script was used to make sure that all the images are healthy and not defected. -
create_data.ipynb: This notebook is used to convert the images into numpy arrays and normalize them, then save the data as a.pklfile. -
train_model.ipynb: This notebook is used to train the face mask detector model. -
utils/model.png: This is an image file that shows architecture of the model. -
utils/MobileNetV2.h5: The model used to make predictions. -
video_cap.py: This Python script uses your webcam to capture video and uses the trained model to detect and highlight faces with and without masks in real-time
-
Real-time face mask detection: Detects whether a person is wearing a face mask or not in real-time video streams.
-
High accuracy: Utilizes a pre-trained deep learning model to achieve reliable results.
-
Easy-to-use: Provides an easy way for running the detection and integrating the model into other projects.
Once the project is set up, you can use the face mask detection feature. You only need two files to run the project; utils/MobileNetV2.h5 and video_cap.py. Here are the main usage instructions:
-
Run the real-time face mask detection script:
python video_cap.py -
A live video stream will open, and the application will start detecting faces and their mask-wearing status.
-
To quit the application, press the 'q' key.