Amrita students Nitin Dantu, Vimalkumar, Dana Vishnu along with their mentor Dr. Shriram K. Vasudevan (Assistant Professor, Department of Computer Science and Engineering, Amrita School of Engineering, Amrita Vishwa Vidyapeetham, Coimbatore campus), developed Diagnostix, an app to classify COVID-19 positive and non-COVID-19 patient X-rays using deep learning algorithms and blockchain.
One of the major symptoms of COVID-19 is the common cold. In the Diagnostix app, there exists a subtle difference in X-ray images of lungs from COVID-19 positive vs COVID-19 negative patients. There is no solution at present that has been able to capitalize on this feature and develop a suitable classifier. Throat swab test is heavily dependent on the medical equipment and the current situation prevailing in the country is not apt for it.
The app is a web/mobile platform with complete diagnostics that allows users to upload X-ray images and the system then predicts. In case the result turns out to be COVID-19 positive, the patient’s details and his/her current location are appended to the blockchain as a transaction for authorities to monitor. They allow us to collect the Aadhar details, phone number, name and GPS coordinates of test location, etc., and for safety and tamperproof storage, are moved to the Blockchain. Once the patient has recovered, the status gets changed as COVID-19 negative and they believe that this as a 100% viable approach. They give 95% accuracy in classifying pneumonia/non-pneumonia and almost 100% accuracy to detect the COVID-19 positive cases.
The app facilitates reduced load on doctors, reduced time of identification of probable candidates, using test kits only on most probable suspects, affordability and easy of use.