Understanding the immediate needs, our application provides a major contribution to these tragic moments. Being alert of viruses and wearing precaution measures is only the option we have right now. The application developed is not intended to recover or completely remove the cases of COVID-19 but it helps to reduce the transmission of the virus from one person to others resulting in a decrease in the number of future cases. It will alert the safety precaution you must take when you are in any kind of unsafe environment. Although “Mask-Shield Analyzer” is now formed particularly to detect the safety measures and reduce the transmission of the virus, we can adopt this application to reduce the transmission of every disease which gets transmitted through respiratory droplets or by the contact of contaminated objects or infected persons themselves.
What it does
In the present situation, the only safety measure taken in our country to detect COVID-19 precaution is checking body temperature and alerting for wearing masks, gloves, and PPE which is done manually. In most public places, the safety rules are written on notice boards which is not so effective in general. Especially during this pandemic, we need digitized real-time alerting and detection systems. If a single individual having an infection is not wearing a mask and other precautions in public places or any sensitive area can infect multiple people at one time. There is no such application that identifies the risk for people in real-time and contributes to minimizing the spread of the COVID-19 virus. Our application provides a major contribution to this issue. It will alert the safety precaution one must take when they are in any kind of public place or sensitive environment.
How we built it
We have used YoloV3 for object detection on our custom trained datasets of around 3500 labelled images.
Challenges we ran into
The major challenge was data set collection. The more labeled images we have, the more accurate our neural networks become. We didn’t find any open source image dataset appropriate as our application requires, so 1 third of our team was working on collecting and labeling images. Second challenge we had was to decide what will be the features we want to build in its first phase(in Garuda Hackathon). We can design the features (increase or decrease numbers of safety kits to be identified from the frame) as the environment and situation needs but due to time constraints we could only work on first few major safety kit identification like mask and face shield.Accomplishments that we’re proud of What we learned
Accomplishments that we're proud of
The application developed will help the community by measuring the safety index through the detection of whether the person in the frame of image or video is wearing minimum precaution or not. The level of precaution that needs to be detected in different places can be customized according to how risky the environment is. This application will detect what are the precaution measures the person in the frame is wearing, and analyze if they are safe enough for the environment
What we learned
Team building and prototyping our idea into reality.
What's next for Mask-Shield Detection System
We have seen the world suffering from various dangerous diseases in earlier periods too like Ebola Virus, Spanish Flu, Influenza, etc which took the lives of many people in the world. It was difficult to control the transmission of those diseases in the early stage which caused great loss to the whole world. Similarly, we can imagine the world will suffer from such pandemic any time in the future too. Understanding the uncertain times, our project will focus on detecting the COVID-19 Safety Kit Analyzer (7 protective kits i.e. mask, gloves,head cover, face shield, goggles, boot cover and PPE). This can be used not only for controlling specific Corona Viruses but can be used to control the transmission of other diseases transmitted through respiratory droplets, air, or the touch of contaminated surfaces or people in future. For the underdeveloped and developing countries like Nepal, the “COVID-19 Safety Kit Analyzer” can be the best digital idea to reduce the spread of such communicable diseases by detecting the level of safety measures a person is adopting.