IISc and Oslo University develop ‘AnamNet’ to assess severity of lung infections in Covid-19 patients

Posted on Updated on

The Indian Institute of Science (IISc), Bengaluru and the Oslo University Hospital and University of Adger, Norway have developed AnamNet, a novel software tool that reveals the severity of lung infections in Covid-19 patients.

AnamNet can look for abnormalities and classify areas of the lung scan as either infected or not. The tool embedded with artificial intelligence can judge the severity of the disease by comparing the extent of infected area with healthy area.

This work in part was supported by the WIPRO GE-CDS Collaborative Laboratory on Artificial Intelligence in Healthcare and Medical Imaging as well as Indo-Norwegian Collaboration in Autonomous Cyber-Physical Systems (INCAPS), INTPART Programme, Research Council of Norway.

The Departments of Computational and Data Science and Instrumentation and Applied Physics at IISc developed AnamNet, using a special kind of neural network that helps to assess the extent of lung damage by searching for specific abnormal features.

“Now this is the first-of-its-kind software tool that can natively run on mobile phone in a standalone mode without utilising any cloud computing. Oslo University Hospital team provided the annotation data required for AI algorithm. Those from University Agder enabled the implementation on other embedded systems like Raspberry Pi and NVIDIA Jetson,” Phaneendra Yalavarthy, Associate Professor, Department of Computational and Data Sciences, IISc told Pharmabiz in an email.

As academic innovators, we work towards open-source and this innovation is publicly available. It is developed with clinicians from Oslo University Hospital radiologists. It is not for diagnostic purpose as of now as it has to get regulatory approvals. We started working on it in March 2020 and the full development took about 6 months, he added.

The findings have also been published in the journal IEEE Transactions on Neural Networks and Learning Systems.

“AnamNet employs deep learning and other image processing techniques, which have now become integral to biomedical research and applications. The software can identify infected areas in a chest CT scan with a high degree of accuracy referred to as anamorphic image processing,” said Naveen Paluru, first author and PhD student in the lab of Phaneendra Yalavarthy, CDS.

“AnamNet extracts feature the CT images of the chest and projects them onto a non-linear space in a mathematical representation to recreate the image from this representation. The study also compared AnamNet’s performance with other advanced software tools which perform similar tasks where it matched its peers in accuracy, to perform just as well using fewer parameters,” stated the researchers.

Moreover the neural network was also computationally less complex, which allowed the researchers to train the tool to detect anomalies faster. A big benefit of AnamNet is that the software is lightweight with a small memory footprint enabling researchers to develop an app ‘CovSeg’ which can operate on a mobile phone and be used by healthcare professionals, they said.

“We felt the need for a lightweight framework that could be deployed as a point-of-care diagnostic device on smartphones or a Raspberry Pi. This feature is missing from currently available technologies like UNet, which requires specialised hardware,” stated Paluru.

The researchers noted that AnamNet holds promise beyond merely identifying lung infections in Covid-19 patients. Efforts are on to look at other common lung diseases like pneumonia, fibrosis, lung cancer and with certain changes to the present design, it could be also used to read brain scans.

Source : 1

Let us know what you think!

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s