Realising the need to develop a cloud based automated diagnostic tool to detect coronavirus disease using radiographs. Adamas University, Elmax Systems & Solutions (ESS) has joined hands to come out with a low cost solution AI enabled radiograph to tackle the problem on Multiple Hypothesis Inference Criterion (MHIC).
The automatic detection tool is said to be unique as it can bring in a single point multi-hypothesis based faster detection of COVID-19 from radiography images. The tool can also detect other vital medical conditions with a severity score tag for early detection of COVID-19 patients and necessary measures.
The two organisations have signed an MoU to offer a valuable second opinion to the doctors and assist them in the screening process.
Informing that this type of mechanism will provide results to the doctors quickly and in an efficient and accurate manner, Debashis Chakrabarti, CEO of ESS and alumni of IIT-Kharagpur said, ” In this software, we have developed an Artificial Intelligence based technique for detecting COVID-19 on Chest Radiographs(X-Ray or CT Scans) ,other medical vitals like temperature , saturation percentage of Oxygen and even pathological data from Blood Routine Examination like Leukocyte, Lymphocyte & Neutrophil.”
He also added that the technique would also act as an aid to the conventional RT-PCR based method to isolate the contracted individuals.
The project will also see utilization of the techniques of data science and machine learning on radiograph images along with other parameters to successfully detect the COVID-19.
The principle-Investigator of the project Dr Moumita Mukherjee, Associate Professor & Dean (R&D) – Adamas University added that, “This software product is built around the same concept with some important additions that can conclude the presence of the stain. Our software can even predict the Stage of the COVID-19 Infection in terms of opacity and lung consolidation. We have achieved 85 percent of accuracy level – we are trying hard to enhance that to something beyond 90 percent, so that the value could be enhanced from the market point of view”.
“The first phase of clinical validation of this software is already done with a large number of radiograph data from normal and affected patients where the team has received more than 80 percent accuracy level and work is going on for further optimization. We are trying to collaborate with government sectors/health-sectors to get more data from COVID hospitals for the large-scale clinical trial of the software. In their words this software eliminates the false negatives and identifies the asymptomatic cases and other cases like Pneumonia, SARS (Severe acute respiratory syndrome) etc. by making use of MHIC . The entire test procedural time is ranged from 15 minutes to 1 hour depending on the modules chosen for the tests. Hence it may be viewed as a faster, lower cost complimentary detection tool for COVID – 19, which can even work on an OPD environment,” Mukherjee added.