AI model of Mask Recognition Project conducted during the "Deep Learning" class offered by Télecom SudParis. The final model contains a real-time face mask detector using Tensorflow, Keras, and OpenCV with your webcam.
The dataset used is provided by Prajna Bhandary (https://github.com/prajnasb/). It contains 1376 images divided into with mask and without mask
- Built using the VGG16 CNN Architecture with layers added on top of the base model through transfer learning.
-
The following results were obtained using our model for artificially generated masks and real masks.
-
Real life examples were made using the faces of our group members to showcase the model's accuracy:
- Built using the MTCNN model for image finding on each frame, a scale-invariant method.
- Below is a link to the video containing a live demonstration of the video results.
- https://www.youtube.com/watch?v=Z5UPBoSzV4s
[
]