Given the current trends in incidence and underlying healthcare system vulnerabilities, Africa is facing a lot of problems due to the Covid-19 pandemic such as a drastic reduction in medical commodities and supplies following border closures and restrictions on exports, and financial resource limitations. A lot of people these days are avoiding the use of basic tenets of hygiene during this crucial time such as wearing masks and gloves in public places. Moreover, they endanger establishments by not abiding by the guidelines and compromising the safety of themselves and others. That is why we came up with the idea of Maskaught, a simple mask detection system that can be placed outside any shop which has access to basic surveillance cameras.
WHAT IT DOES
Maskaught is a convenient mask detection that system that can be placed outside any shop or establishment. This system would ensure that those who are found to be not wearing a mask would not be allowed to enter inside. Outside a mid-range or large scale shop with security, it can act as a helping hand to the security guard by minimizing his/her interaction with consumers with the help of this software and ensure their safety.
HOW WE BUILT IT
We first trained the system by providing a dataset containing pictures of people with and without masks. After training the system, all the images were converted into an array and two deep learning models for detecting face and mask were created. The accuracy of testing was also plotted as shown in the graph. Both the models were then loaded into a new Python file and a camera was integrated into the mask detection system. The system would then detect the mask on the face and displays the accuracy of detection. A text to speech software was also integrated within the system which would guide the customer throughout the whole process of detection.
Converting our python project into a web application was a significant challenge that was faced by our team. However, we used python's Django framework to bridge the gap between our python software and HTML. As a result, we were able to build an interactive and user-friendly interface for the project.
Completing a challenge always feels satisfactory. Thus the entire project from the mask detection to the web application to our business model are all accomplishments that we are proud of.
WHAT WE LEARNED
Through this hackathon, we had the opportunity to learn how to train a deep learning model and create a python program integrating the use of Keras, OpenCV, and MobileNet along with text to speech conversion software (pyttsx3). We also applied our web development skills to attach the python program to the HTML one using the Django framework.
Our idea also focuses on promoting the rise of small domestic businesses that do not get a lot of customers as they’re not able to keep a track of whether all their customers are wearing masks inside. This may sometimes procure them with financial losses as it may cause customers to stop coming to these shops due to the ease of contracting the coronavirus with people without masks. Moreover, since our idea is inexpensive to enact where we need to only connect our web app to the camera the cost involved to adopt this system is pretty much minimal. One more plus point regarding our system is its expanded use and modifications which can be made after adopting it. It can be further applied for security in the future when we are safer against the pandemic. It can also add in more scanning features in the future like scanning for gloves. With further improvements, this system could be integrated with CCTV cameras to detect and identify people without masks and could be used in the imposition of fines for people who don’t wear masks by the government.