The COVID-19 pandemic is a threat to public health and safety, and it is critical to maintain social distancing and wear face masks to control it.
What it does
To remind people to wear masks, our group built a program to capture people’s faces in real time and check if they wear a mask or not.
How we built it
We started our project from learning and using the cv2 library, including taking in images from our laptop camera and capturing faces using the built-in Cascade Classifier. Then, we learned the basic model of neural network and constructed a simple convolutional neural net using layers from the Pytorch library. After training our model to reach a desired accuracy, we combined the facial recognition and neural net together to display the mask wearing status when a face was recognized.
Challenges we ran into
Neural nets not updating correctly Overfitting occurred when learning rate/epoch number is too high Laptop camera responded slowly during facial recognition
Accomplishments that we're proud of
We've accomplished it!!!
What we learned
How to use opencv for facial recognition How to build a neural network How to display words on real time video
What's next for Mask Recognition
For future extensions, we can develop the neural net to reach a higher accuracy in facial or mask recognition. Additionally, this program can be deployed among mobile robots to monitor mask wearing status in public spaces and better control the pandemic.