Researchers at the Indian Institute of Science (IISc) have designed a special tool called “AnamNet” to diagnose the severity of coronavirus infection in the lungs of patients.
The study was carried out by researchers from the departments of Computing and Data Science (CDS) and Instrumentation and Applied Physics at IISc. They conducted the study in collaboration with colleagues from the University Hospital in Oslo and the University of Agder in Norway.
The study was published in the journal IEEE Transactions on Neural Networks and Learning Systems.
To develop this software tool, the researchers used a unique neural network and deep learning, along with other image processing techniques. Their tool finds specific abnormal features and estimates damage in the lungs and identifies infected areas in a chest CT scan with a high level of accuracy.
The authors noted that the app’s neural network was also less computationally complex, allowing researchers to train it much faster to detect anomalies. Being lightweight software, AnamNet leaves only a small memory footprint.
The software is designed to âsegmentâ the scan as âinfectedâ and âuninfectedâ.
The researchers explained in the study that the tool essentially extracts features from chest CT images and projects them onto a non-linear space, then recreates the [segmented] image of this representation.
The researchers said they plan to expand the software tool’s services to detect and trace other lung diseases, including pneumonia, fibrosis and even lung cancer.