A new system for analyzing chest CT scans with deep learning enables detection of COVID-19 lesions

Credit: CC0 public domain

A new automated system that includes deep learning technology will enable the detection of COVID-19 lesions by analysis of computed tomography (CT) scans.This system is described in a study published in the journal Biology and medicine computerWas conducted by researchers at the UB, the EURECAT Technology Center in Catalonia, and the Computing Vision Center (CVC).

This study was able to verify the efficiency of the system as a support tool for medical professional decision-making in the COVID-19 detection task and for measuring gravity, spread, and progression of pneumonia caused by SARS. CoV-2 is in the medium to long term. ” Giuseppe Pezzano, the principal investigator of this study and the investigator of the UB and EURECAT Digital Health Units, said.

Specifically, the function of the system consists of “the first phase of lung segmentation with CT scans to reduce the search area”. “Then algorithm It is used to analyze the lung area and detect the presence of COVID-19. If the test is positive, the image is processed to identify areas affected by the disease, “he adds.

The algorithm has been tested on 79 volumes and 110 sections of CT that detected COVID-19 infection, obtained from three open access image repositories.Researchers have achieved average accuracy in segmentation Lesion Caused by about 99% of the viruses, no false positives were observed during identification.

This method uses an innovative method for calculating the mask for segmentation of medical images and has yielded good results in the segmentation of nodules in tomographic images.

Several recently published studies show that “deep learning and computing vision algorithms have achieved better accuracy than expert cancer detection, stroke and heart attack predictions on mathematics.” , Professor Petia Radeva of the Faculty of Mathematics and Computer Science said. UB. We couldn’t be left behind, so we’re working on this technology and high precision to analyze medical images in an objective, transparent and robust way to help doctors fight COVID-19. Provided the data. Senior Researcher at Computer Vision and UB Machine Learning and Computing Vision Center in the Integrated Research Group.

“This type of automation system is an important tool for medical professionals to make more robust and accurate diagnoses because it can provide information that humans cannot measure,” said Oliver Diaz, a lecturer in the Faculty of Mathematics and Computer Science. Emphasizes. UB.

Vicent Ribas, Head of Research Lines for Medicine Data Analytics in the EURECAT Digital Health Unit, said: The use of artificial intelligence is becoming more useful. ”

Artificial intelligence can speed up stroke detection

For more information:
Giuseppe Pezzano et al, CoLe-CNN +: Convolutional Neural Network for Context Learning-COVID-19-Ground-Glass-Opacities Detection and Segmentation, Biology and medicine computer (2021). DOI: 10.1016 / j.compbiomed.2021.104689

Quote: COVID obtained on December 1, 2021 from https: // by a new system for analyzing chest CT scans with deep learning -19 lesion detection (2021, December 1st) will be possible enable.html

This document is subject to copyright. No part may be reproduced without written permission, except for fair transactions for personal investigation or research purposes. Content is provided for informational purposes only.

A new system for analyzing chest CT scans with deep learning enables detection of COVID-19 lesions

Source link A new system for analyzing chest CT scans with deep learning enables detection of COVID-19 lesions

Show More

Related Articles

Back to top button