International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 11 Issue: 04 | Apr 2024
p-ISSN: 2395-0072
www.irjet.net
ENSURING EDUCATION FOR POST COVID USING DEEP LEARNING Y. Angeline Gnana Prakasi1, A. Nisha Devi2, V. Nivetha3, M. Subha Mughisha4, K. Varushapriya5 1Professor ,Dept. of information technology, Meenakshi college of engineering, Tamil Nadu, India
2,3,4,5 UG Student, Dept. of information technology, Meenakshi college of engineering, Tamil Nadu, India
---------------------------------------------------------------------***---------------------------------------------------------------------based on their unique facial features, eliminating the need Abstract - A deep learning based study was used to reduce
for manual attendance taking methods which could also vastly reduce the problem with proxy attendance.
the COVID effect on education system. Corona virus was identified in 2019 and posed a great challenge to the world since its outbreak. despite of the comment on end of COVID or end of mask wearing era still a report says that there are new 15,000 fresh cases of COVID from January 2024. Even a scientist molly smith stated on wearing mask in 2024 that at certain situation coronavirus epidemics have forced people to wear masks to counteract the transmission of virus, which also brings difficulties to monitor large groups of people wearing masks. In this paper, we primarily focus on the AI techniques of masked facial detection and related datasets by using CNN based deep learning algorithm. The purpose of the proposed system is to protect the students and staff against infectious diseases and increase the student performance during classes by monitoring their facial movements and alert them during drowsy times. The paper also works on detecting the use of masks in closed areas like classroom by training a customized deep learning model which also monitors the student’s attendance data just by recognizing their face through a deep learning model.
The libraries like shutil and pickle could be used to manage the data of the student in the backend of the system. Already enrolled student’s data along with their unique id were refurbished based on their daily attendance. In case of new students, the student is asked to show their face in the camera to train the system freshly. Also the student’s attentiveness in the classroom is monitored by tracking their eye movements, which could improvise the productivity of education.
2. EXISTING SYSTEM The existing system is the first survey about masked face detection dataset. It works on numerous datasets like around 13 Datasets and compare those methods to ensure the best out of them. Initially they are using two categories of methods. (i) Conventional method.
Key Words: convolutional neural network, automatic attendance, deep learning, masked face , drowsiness alert.
(ii) Neural network method. The conventional method usually works on hand crafted features, which accounts for a small proportion and the neural network based method are further classified as three parts according to processing stages.
1.INTRODUCTION After the attack of COVID a lot has changed in the society especially a diverse effect where caused in the education sector. When the situation turns normal, where a one to one classes started again the students and staffs were insisted to wear face masks and later on which become a mandatory one. Hence in our project a 3 stages of initial test were proposed to examine the student’s state and further a smart attendance could also be done by storing and reporting the student data by facial recognition. Also the student’s facial and eye movements were tracked during the class hours to monitor their attentiveness. It is also designed to alert them in case of remissness.
They have also incorporated ACS (Advanced Card Systems) card reader to register the unknown students to the system. Additionally, an MQTT (Message Querying Telemetry Transport) broker that is run on the device establishes the data transmission. 2.1 Disadvantages of existing system
Initially the face of the student is analyzed and features are extracted to create a ROI. Next, the student’s face is differentiated from the mask by using a system, which is already trained using the datasets uploaded. After the advent of COVID, there were many datasets developed to work on the project. Further using facial recognition technology, the system can accurately identify individuals
© 2024, IRJET
|
Impact Factor value: 8.226
|
Here some methods are used to detect only masked and un masked faces and no other advanced features
Some datasets are created by stimulating masks which could not give better result
Lacking of the model size
ISO 9001:2008 Certified Journal
|
Page 1827