Real Time Facial Mask Detection Using Machine Learning

Project Title

Real Time Facial Mask Detection Using Machine Learning

Project Area of Specialization Artificial IntelligenceProject Summary

COVID-19 is a kind of coronavirus that has made its way into our country Pakistan made the government to impose large-scale social restrictions and lockdown to stop the coronavirus from spreading. As the number of patients grew, the administration continues to appeal public to use masks and later on government made it compulsory to wear face mask regardless of their health condition. This appeal is in line with the World Health Organization's (WHO) recommendations for limiting the spread of COVID19. As a result, a system to track those who haven't worn masks in public service locations in real time are needed. 

Project Objectives

1. To learn and gain knowledge about the critical application of Artificial Intelligence system.

2. To learn and gain knowledge about how computer vision, deep learning, and deep learning algorithms like CNN etc. and how they can be implemented with Artificial Intelligence system.

3. To develop a model for face mask detection and warning system using Jetson kit.

4. To training our model in a suitable neural network so as to acquire perfect outcomes

Project Implementation Method

We will create a mask recognition programme with a camera that can be used as a picture and video input and is connected to the Jetson Nano kit, which processes the data and displays it on the LCD.

Benefits of the Project

We created the tools to address issues with persons who did not use masks or who did not wear them so this device is quick enough to warn the concern authority about such person to prevent the spread of COVID-19 Or other air transmitted infections. 

Technical Details of Final Deliverable

Mask detection system is realized based on YOLOv3-tiny, which is small, fast and suitable for mobile hardware deployment and real-time detection. Due to the urgency of controlling COVID-19, the application value and importance of real-time mask detection are increasing. To address this issue, we built the dataset with 7633 quality masked face images and developed Tiny YOLOv3, a fast and accurate mask detector with a channel attention mechanism that enhanced the feature extraction capability of the backbone network. Furthermore, we used GIoU and focal loss and adopted the corresponding data augmentation to improve the accuracy and robustness of the model.

Final Deliverable of the Project HW/SW integrated systemCore Industry HealthOther Industries Medical Core Technology Artificial Intelligence(AI)Other Technologies Wearables and ImplantablesSustainable Development Goals Good Health and Well-Being for PeopleRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	27000
Jetson Nano 4GB	Equipment	1	20000	20000
Camera	Equipment	1	2500	2500
Monitor	Equipment	1	4500	4500