In my community, the number of cases of covid19 is increasing more and more because of the non-respect of the barrier measures given by the government and the United Nations, such as for example the refusal to wear a face mask in public. In order to reduce the spread of covid19. So I got the idea to design a face mask recognition system that will identify people who are not wearing a face mask. it can be integrated with the Automatic Door Opener system of hotels, shopping centers, etc so that the door can only be opened if and only the person wearing a face mask.
What it does
The system takes a photo of a person and then identifies whether that person is wearing a face mask or not
How I built it
I use PyTorch and the transfer learning technique to train the model and thus achieve an accuracy of over 80%.
Challenges I ran into
The challenge was to find an environment to work with the GPU to be able to train the model faster.
Accomplishments that I'm proud of
I am proud to have over 80% accuracy and the prediction is excellent.
What I learned
I learned to use the VGG16 model for the transfer of learning in addition to other notions that I acquired.
What's next for Face Mask recognition with CNN
Future work will be integrated into the automatic door opening system of hotels, shopping centers, etc. so that the door can only be opened if and only if the person is wearing a face mask.