Downloads: 1 | Views: 329 | Weekly Hits: ⮙1 | Monthly Hits: ⮙1
Working Project | Biomedical Instrumentation | India | Volume 11 Issue 7, July 2022 | Popularity: 5.1 / 10
COVID-19 Lung Infection Segmentations using Neural Networks based on CT Slices
Priya Ranjan Dash, Bijay Kumar Ekka
Abstract: Beginning in early 2020, COVID-19, a health emergency and existential threat to society, began to spread globally. The current healthcare approach to combat COVID-19 has the potential to be considerably improved by automated lung infection detection utilising computed tomography (CT) images. To segment contaminated areas from CT slices, however, there are a number of difficulties, including low intensity and a wide range of infectious features when comparing infected and healthy tissues. Additionally, it is not feasible to collect a big amount of data quickly, which inhibits the deep model from being trained. To overcome these difficulties, a novel COVID-19 Lung Infection Deep Network Segmentation is proposed to autonomously separate unhealthy areas from slices of a chest CT image. This study provides a segmentation technique for Ground Glass Opacity or ROI identification in CT images caused by corona viruses. The area of interest was categorised down to the pixel level using a modified Unet model structure. Instead of the time-consuming RT-PCR test, CT scans can be utilised to diagnose COVID-19. Using this segmentation method, doctors were able to diagnose COVID-19 more quickly, precisely, and consistently.
Keywords: CT scan, RT-PCR, ROI identification, Unet
Edition: Volume 11 Issue 7, July 2022
Pages: 1650 - 1655
DOI: https://www.doi.org/10.21275/SR22723205035
Please Disable the Pop-Up Blocker of Web Browser
Verification Code will appear in 2 Seconds ... Wait