Experts study advanced AI nCov scanning

Yagnesh Bharat Mehta | TNN | Apr 21, 2020, 04:28 IST

Surat: With an aim to prevent doctors from getting infected with coronavirus, a mother-daughter medico-technical team studied the AI-based Covid scanning to identify infection in a patient at the preliminary stage without any physical intervention of a doctor.
In their study, the duo developed a concept that it can be verified at initial level if a person has infection through visual examination of throat by an automated AI scanner. The Study aims at prevent doctors who are already limited in numbers from getting infected from exposure to coronavirus during treatment of Covid positive patients.
Dr Upena Dalal, professor in electronics department at Sardar Vallabhabhai National Institute of Technology (SVNIT) and her daughter Dr Parima Dalal Parikh, a paediatrician, undertook a study of throat infections in Covid-19 patients and developed a method to examine it electronically by a AI-based scanner. The duo took a lead from Krishna Ojha and Ekta Arora, two BTech students of Indian Institute of Information Technology (IIIT), Surat, who developed an initial model of AI based Covid19 scanner.
“The students’ AI-based Covid-19 scanner notices basic symptoms of sneezing, coughing, runny nose and chest X-ray. But chest infection is developed at a later stage and in that condition a patient cannot even stand properly,” explained Dr Parikh.
But during their study, the duo found that tonsillitis with some specific characteristics can suggest if it is due to coronavirus.
“After studying the findings of doctors and examination details of coronavirus at the initial infection stage, we arrived at a conclusion that if the AI-based scanner can notice and identify the infected throat, it will be more effective for treating it at once,” said Dr Dalal.
The mother-daughter team has also shared their findings with the two IIIT students. “Now we are working on use of infra-red thermal and optical camera to examine throat infections. With this, a patient can be examined by the scanner without any doctor coming in direct contact with the infected person,” Dr Dalal added.
However, the study is still on the theory table without having conducted any clinical validation as yet.
The team is now working on to get support from the MHRD to develop the project at an advance level which can be used in hospitals and health centres.