Mahesh Tunguturi and Samrajyam Singu; 2 Software Engineers Who Are Developing AI to Determine the Severity of Covid-19

In no sector is this involvement more vital than in health, where the Covid-19 pandemic has been a catalyst for new AI technology that will help support Covid-19 patients. Scientists Mahesh Tunguturi and Samrajyam Singu are some of the innovators at the leading edge of this AI research with their latest project.

As hospitals (particularly intensive care units) struggle to support the number of Covid-19 patients, tools are being developed to assist, speed up and simplify the diagnosis and treatment of patients. One such tool is an AI system that can be used to identify the severity of Covid-19, as well as help predict which patients are likely to suffer from further complications.

This AI tool was created by Indian-born software engineers Mahesh Tunguturi and Samrajyam Singu who are in the midst of developing an algorithm that can read and assess chest x-rays. The invention is designed to substantially reduce each Covid patient's need for full-time access to a radiology specialist, something that is hard to come by during the pandemic.

Tunguturi and Singu detail in their research how AI can use image quality, digital image processing, and segmentation and classification of the lung tissues to assess what is known as the affectation index (IAF). The IAF is used to determine a patient's current diffuse lung border condition and can mark patients as severe or critical. As Covid patients frequently suffer from pulmonary complications such as pneumonia, this technology has the capacity to help diagnose conditions for earlier treatment or prevention than was previously possible.

The proposed process outlined by Tunguturi and Singu begins with image quality evaluation, where contrast, blur, noise, artefacts, and distortion are all assessed to ensure important details on the x-rays are not lost. Then, digital image processing is used to prepare the images for screening by using a filter that enhances their quality by filtering high frequencies. After that, a segmentation process will be used to section parts of the lung in order to better assess each region. Finally, the pixels on each image are used to classify the lung tissue and create the IAF.

According to Tunguturi and Singu's research, the resulting IAF can then be used to assess a patient's lung condition and thus provide important information like how progressed a patient's Covid symptoms are, whether they have any additional pulmonary complications and if they are at risk of developing lung issues at a later time.

Tunguturi and Singu believe that this system can reduce the need for full-time radiological support for Covid patients, as well as accurately estimate a patient's illness severity; thus alleviating pressure on healthcare systems.