Abstract
COVID-19 is a disease which spreads with human-to-human interaction. In this situation, humans need to be distanced from each other. To prevent the pandemic, people need to wear masks. Wearing a mask is mandatory for all people. That's why it is important to detect whether a person is wearing it or not. This chapter aims to provide a biometric approach for COVID 19 prevention. In this research the machine is able to check whether a person is wearing a mask or not from an image or a live stream. This research is also a smart approach for smart cities. In this research, the authors generate the artificial dataset by adding the mask on the faces of the persons in the images. And that can be done by the automation of the unmasked people's images dataset. The dataset is trained using tensor flow/keras for providing the classifier which classifies the image. This research is valid on image and on live streams.
Top1 Introduction
Face recongination technique is a very popular technique in this world of artificial intelligence. Through this technique, one's face is recognized, due to which many problems can be solved, this is a face mask detection based on which the face is detected. Facial recognition techniques include a user-friendly system, noncontact, high concurrency. The facial recognition system has large applications such as security, retailing, e-commerce, etc. (Firdaus F. et al., 2022).
Facial recognition by ai which helps to detect a face by every part and every corner of a face is also used in phone laptops and many electronics devices. AI facial recognition is used in phones for unlocking phones and also used in laptops so that we can open the laptop by our face. Facial recognition also provides additional security to a system for example: laptop, phones, Tablet, etc. (Chawla D., et.al., 2018) (as per Figure 1).
1.1 Phases of Face Mask Detection
Face mask detection is divided into two phases to detect the ace mask of the user. These two phases are - train face mask detector and apply face mask detector (Bhati D., et.al. 2017).
Train face mask detector is the phase where the detector plays three roles . First the dataset masked and unmasked both is loaded and then the dataset is trained using tensor flow/keras. The last step of this phase is serializing the classifier into the disk (Rosebrock, A. et al., 2020).
Applying a face mask detector is the phase where the main process is done. It consists of so many processes. First, as all the classifiers are stored in disk so it loads from the disk and then detects the image or stream. There are many objects in each image rather than face to face so it is necessary to first extract the ROI (Reign of Interest). After extraction the classifier checks that a person is wearing a mask or not. Classifier results the status of the person (Bhadauria R. V. S., et.al. 2022) (as per Figure 2).
1.2 Facial Recognition as Biometric Security
In today's time, biometric is being used for the security of any important document or files or for attendance in the office. Technologies like fingerprints are being used. Which is not giving that much security now, people are using it wrongly. That's why face security is being used the most in the present time. Through this, important documents or files can be kept more secure, in which face recognition software is used.In which the user's face is first recognized, then the match is done with the stored data set, thus the face recognition software provides a lot of security. Facial recognition is also very important for home security (J. Pedraza.,et,al 2010).
It is shown in this image that human people can keep their important files and mobile secure through face lock. Detect the face by using deep learning algorithm (Gupta K., et.al., 2020) (as per Figure 3).
Key Terms in this Chapter
R-NN: RNN stands for Recurrent Neural Network, this is a type of artificial neural network that can process sequential data, recognize patterns and predict the final output. This is called recurrent because it can repeatedly perform the same task or operation on a sequence of inputs.
CNN: CNN stands for Convolution neural networks. CNN have an input layer, an output layer, numerous hidden layers and millions of parameters, allowing them to learn complicated objects and patterns.
PyTorch: PyTorch is well supported on major cloud platforms, providing frictionless development and easy scaling. Install PyTorch Select your preferences and run the install command.
Classifier: It is an algorithm that is used to map the input data to a specific category.
MobileNetV2: is a classification model (distinct from MobileNetSSDv2) developed by Google. It provides real-time classification capabilities under computing constraints in devices like smartphones.
OpenCV: is a huge open-source library for computer vision, machine learning, and image processing. OpenCV supports a wide variety of programming languages like Python, C++, Java, etc. It can process images and videos to identify objects, faces, or even the handwriting of a human.