World-first computing and artificial intelligence COVID-19 severity scoring study

8 MAY 2020

World-first computing and artificial intelligence COVID-19 severity scoring study

A new research project funded by Charles Sturt University will examine how computer deep learning and artificial intelligence can detect COVID-19.

  • Computer deep learning models with novel image processing techniques will aim to detect COVID-19 in patients
  • Multi-modal pilot study is thought to be the first of its kind in the world

A new research project funded by Charles Sturt University will examine how computer deep learning and artificial intelligence (AI) can detect COVID-19.

The study will use deep learning models with novel image processing techniques, which will be capable of detecting COVID-19, to estimate a corona severity score to help screen critically ill patients.

Lead researcher Professor in Computer Science Manoranjan Paul in the Charles Sturt School of Computing and Mathematics said he believes this multi-modal pilot study will be the first of its kind in the world.

“The project aims to develop efficient deep learning models with novel image processing techniques, which will be capable of detecting COVID-19,” Professor Paul said.

“We will then be able to estimate the corona severity score of this disease in terms of the Australian Triage Scale (ATS) utilising multi-modal data.

“The outcome of the project will support medical professionals in COVID-19 triage to identify and monitor the progression of this disease in multiple patients in parallel, based on multi-modal data analysis and deep learning.

“The proposed system will provide accurate diagnosis, prognosis and screening of patients who can be flagged for further review by a radiologist or clinician for possible treatment/quarantine.”

The proposed system will be trained on a range of available imaging modalities (like X-ray and chest CT), age, morbidities, and breathing patterns.

The project will also investigate imaging features that lead to death.

A scoring system based on the available data will be developed to screen critically ill patients who need immediate medical attention.

“The proposed project is based on our recent preliminary studies (see Media Note below) where we have already proved that deep learning and AI can detect COVID-19 from pneumonia using lung X-ray with 88 per cent accuracy,” Professor Paul explained.

“This work will generate new knowledge regarding the impact of COVID-19 on our people or communities by evaluating a disease progression monitoring framework.

“To the best of our knowledge, this multi-modal pilot study will be the first of its kind in the world, and the findings from the study will lead to a more rigorous study in near future.”

Professor Paul expects the $30,000 research pilot study will be completed by the end of 2020.

The Charles Sturt research project team is:

Lead researcher Professor Manoranjan Paul, with co-researchers Dr Anwaar Ulhaq (Charles Sturt University), Dr Subrata Chakraborty (Univerisity of Technology Sydney), Dr Manash Saha (Manning Rural Referral Hospital, Taree NSW), Dr DM Motiur Rahaman (Charles Sturt University), and Dr Tanmoy Debnath (Charles Sturt University).


Media Note:

To arrange interviews with Professor Manoranjan Paul contact Bruce Andrews at Charles Sturt Media on mobile 0418 669 362 or via news@csu.edu.au

The recent preliminary studies are:

[1] Ulhaq, A, Khan, A, Gomes, DPS, and Paul, M, (2020). Computer vision for COVID-19 control: A survey. https://doi.org/10.31224/osf.io/yt9sx, 

[2] Horry, M J, Chakraborty, S, Paul, M, Ulhaq, A, Pradhan, B, Saha, M, and Shukla, N (n.d.). (2020). X-Ray Image based COVID-19 detection using pre-trained deep learning models. https://engrxiv.org/wx89s/


Share this article
share

Share on Facebook Share
Share on Twitter Tweet
Share by Email Email
Share on LinkedIn Share
Print this page Print

All Local News Computer Science Mathematics Medicine