Step-by-Step Implementation

The Face Mask Detection model is created in four steps:

1. Specifying the model : (layer node, the activation function is applied to those nodes)
2. Compile : (loss function, Optimizer)
3. Fit : (make model learn)
4. Predict : (use the model to predict)

To train a customized face mask detector, we must divide our project into two unique stages, each with its own set of sub-steps (as seen in Figure below):

Training: Here we’ll focus on loading our face mask detection dataset from disk, training a model (using Keras/Tensor Flow) on this dataset, and then serializing the face mask detector to disk.

Deployment: Once the face mask detector is trained, we can then move on to loading the mask detector, performing face detection, and then classifying each face as with _mask or without_ mask.

Model Description

1. Face Recognition
2. Face Mask Detection
3. Temperature Check
4. Alert System & Hand Sanitizer

1) Face Recognition:

Face detection is a sort of computer vision technology that can recognize people’s faces in digital photographs. 

• Facial recognition entails recognizing the face in a picture as belonging to person X rather than person Y. It is frequently used for biometric applications, like unlocking a smartphone. 
• Facial analysis attempts to learn something about people based on their facial features, such as their age, gender, or the emotion they are displaying.
• Facial tracking technique is commonly used in video analysis and attempts to follow a face and its features (eyes, nose, and lips) from frame to frame.

2) Face Mask Detection:

Data At Source: OpenCV was used to increase the size of the images. At the time, the images were titled “cover” and “no veil.” The images available were of various sizes and goals and were most likely extracted from various sources or from machines (cameras) of various goals.

Data Processing: Ventures, as indicated below, were applied to all the raw data images to convert them into clean forms that could be handled by a neural organization AI model.

• Resizing the information picture (256 x 256).
• Applying the shading sifting (RGB) over the channels (Our model MobileNetV2 underpins 2D 3 channel picture).
• Scaling/Normalizing pictures utilizing the standard mean of PyTorch work in loads.
• Center trimming the picture with the pixel estimation of 224x224x3.
• Finally Converting them into tensors (Similar to Numpy exhibit).
• Training and,
• Deployment.

3) Temperature Check:

Although no thermal cameras can detect or diagnose the coronavirus, FLIR cameras can be used as an addition to existing body temperature screening methods in high-traffic public venues to identify higher skin temperature through swift individual screening. If the temperature of the skin in crucial places (particularly the corner of the eye and the forehead) is higher than normal, the individual may be chosen for extra screening.

4) Alert System & Hand Sanitizer:

The buzzer and sprinkler motor will be linked to the microcontroller. If the temperature threshold is exceeded and the face mask is not detected, the alarm will be activated. If both the face mask and the temperature are set to normal, the automatic sanitizer will activate.