Logotyp KTH Logotyp Karolinska Institutet Logotyp Region Stockholm

Project Improved diagnosis of Covid-19 with computed tomography and AI image analysis

Research leader Mats Persson, KTH
3D rendering of computed tomography image of the lungs of a covid-19 patient. The green solid objects are infected lung tissue that has been marked with the help of an image analysis software developed within the research collaboration between KTH and Karolinska University Hospital, and which can speed up the analysis of lung changes in X-rays.

Lung X-rays and computed tomography play a major role in the care of covid-19 patients. At the same time, knowledge on how X-ray images should be interpreted is still limited. The departments of Medical Technology and Medical Image Physics at KTH have in collaboration with Karolinska University Hospital developed a software that largely automates the analysis of X-ray images of covid patients, by marking the lungs and lung damage caused by covid-19. This can save huge amounts of time for physicians who analyze large amounts of patient images to find patterns that can help with treatment. MedTechLabs also develops AI tools based on deep neural networks to automatically analyze computed tomography images of covid patients and predict the patient’s future course of the disease.