• Amidst "Corona Time," when everyone was forced to wear a mask, all the existing face detection and recognition models became useless. To overcome this situation, I have built a mask detector and a masked face recognizer to help the pace of society.
• I have constructed and trained the model using Tensorflow, Keras, VGGNET-16 Transfer Learning, OpenCV, and CaffeModels.
• I have prepared my own database containing the faces of 12 people, and I have also made a separate program to process the data for usage.
• Achieved an accuracy of 85% on testing data and an accuracy of 98% on training data.
• I have also prepared a program to see the recognition in real time.
I've created two files ('FileToSeeTraining.ipynb' and 'FileToSeeRealTimeResult.ipynb') for your convenience to see the training process and real-time recognition.
Please see the drive link below to access the dataset and model files (.h5) I used for this model.
Link - https://drive.google.com/drive/folders/1Bf_CcX3HkLmNBoUiu5NLhNnnLkt1k4z0?usp=sharing
You are asked to make improvements to the model or any other feature.