Automated detection of lung pathologies using ultrasound imaging

The objective of this project is to automate the interpretation of lung ultrasound images to detect artifacts associated with pneumonia. The lung ultrasound scans could be used in point-of-care units to detect and monitor pneumonia in severe COVID-19 patients. The studies have shown that lung ultrasound scans offer a more accurate diagnosis of interstitial pneumonia in comparison to radiographs. Although computed tomography (CT) scans provide better quality images to detect lung pneumonia, repeated CT scans represent a high radiation dose. The Radiological Society of North America (RSNA), Society of Thoracic Radiology and American College of Radiology (ACR) do not recommend using CT scans for routine screening to identify pneumonia in COVID-19 patients. In contrast, ultrasound imaging is ionizing radiation-free and inexpensive. Due to their small size, portable ultrasound scanners are relatively easy to disinfect. However, the interpretation of the lung ultrasound scans requires a significant amount of time as these scans are typically saved as videos. The recent advancements in medical image processing and machine learning techniques offer an excellent solution to automate the interpretation to reduce the time spent on the diagnosis.

Faculty Supervisor:

Kumaradevan Punithakumar

Student:

Partner:

MEDO.ai

Discipline:

Computer science

Sector:

Agriculture; Professional, scientific and technical services

University:

University of Alberta

Program: