Pregledni rad
https://doi.org/10.55378/rv.49.1.6
Advantages and limitations of artificial intelligence tools in the analysis of thoracic computed tomography images in patients with COVID-19 pneumonia – a review
Nina Mrvoš
; Faculty of Health Studies, University of Rijeka, Rijeka, Croatia
*
Melita Kukuljan
; Faculty of Health Studies, University of Rijeka, Rijeka, Croatia; Clinical Hospital Center Rijeka, Department of Diagnostic and Interventional Radiology, Rijeka, Croatia
Maja Karić
; Faculty of Health Studies, University of Rijeka, Rijeka, Croatia; Clinical Hospital Center Rijeka, Department of Diagnostic and Interventional Radiology, Rijeka, Croatia
Klaudija Višković
; Faculty of Health Studies, University of Rijeka, Rijeka, Croatia; University Hospital for Infectious Diseases „Dr. Fran Mihaljević“, Department of Radiology and Ultrasound, Zagreb, Croatia
* Dopisni autor.
Sažetak
Coronavirus disease 2019 (COVID-19) is a systemic viral infection caused by the highly infectious SARS-CoV-2 virus, which can lead to pneumonia and other complications. It first appeared at the end of 2019 in China and spread rapidly around the world causing authorities to declare a pandemic at the start of 2020. People infected with this virus usually present with symptoms such as cough, fever and shortness of breath. This respiratory disease causes changes in lung parenchyma that can be analysed on a CT scan. CT became one of the most accurate diagnostic methods for confirmation of COVID-19 in patients as well as a tool for estimating the stage of infection. Since identifying the disease from radiological scans is quite time-consuming and complex, the development of artificial intelligence (AI) models has emerged as a potential solution to enhance diagnostic speed and accuracy. Most common AI models were based on deep learning and machine learning and were able to correctly identify COVID-19 pneumonia on CT scans. The main problem in developing these models proved to be the lack of databases on which these models could be trained. This review analysed the results of twenty-one papers published on the topic of AI use in detecting COVID-19 on CT scans. The results of analysis show that AI is significantly accurate in detecting COVID-19 pneumonia, but it still needs to be developed further.
Ključne riječi
Hrčak ID:
330411
URI
Datum izdavanja:
11.4.2025.
Posjeta: 119 *