Circumstances are such that we are all familiar with words like RT-PCR and antigens. Though the current COVID-19 tests may be certified, there is no doubt that a more rapid, cheap and effective way of testing needs to be devised. And fast. In this context, meet EuroForce-IT. They have been using the power of robotics, AI and other cutting-edge technology to solve problems. And given the current situation, they have a plan. "We have come up with an AI-based prototype that can record a patient's coughing pattern through our app and predict if they are COVID positive or not," explains Co-founder Rakesh Kumar Sahu. Can you imagine the possibility? One doesn't even have to step out of the house, results are generated immediately, there is no kit required, testing can be done at scale and it's cost-effective.
But alas, EuroForce-IT's path is riddled with one major obstacle — procuring the cough data. Only after they collect enough COVID-19 patient cough data, about 1,000 samples to be exact, can they start the process of feeding it into their prototype. The more the data, the more they can optimise their model. "When we visited hospitals in Odisha in April to ask for this data, they did not even let us in, stating that this is private data and they cannot share it," says the Khordha-based entrepreneur. Currently, they do have recordings of coughs from friends and are waiting for privacy clearance to procure more samples.
Meanwhile, the start-up, founded in 2017 and currently incubated at KIIT-TBI, experimented with data of children's cries. They analysed the recordings of children crying and their AI model could predict why they're crying — out of hunger, sleeplessness or other discomforts. It's the same model they want to apply for COVID testing as well. "But we are facing two problems. Some doctors don't understand the scope of this technology and what we are proposing to do, while a few others think we are trying to take their jobs away," says the co-founder who pursued his BTech in Electronics and Communication Engineering from Biju Patnaik University of Technology.
The team estimates that, if they successfully get the data, they will be in a position to deliver their model in about two months. "India is a large country. Getting the population tested is like asking them to vote, it requires a massive amounts of planning and investment and still, only 60 to 70 per cent will turn up. Hence, in a situation like COVID testing, this prediction model is a better solution," explains Rakesh.
This is the plan to execute their model -
- Collect cough data of positive and negative patients
- Test hypothesis
- Some funding to execute at scale
- Lunching it as a smartphone app
For more on them, check out euroforce-it.com
