EN | عربي

Disease Detection AI

Using artificial intelligence (AI) for accurate pneumonia detection
Play Video

Project Description



Principal Investigator

Xin Gao


Pneumonia causes 2.5 million deaths worldwide every year. Over 50% of those who die are aged 70 or older. Pneumonia also accounts for 15% of all child deaths – making pneumonia the leading cause of childhood death.
There are more than 30 types of pneumonia that are often hard for doctors to distinguish between. This difficulty leads to mistreatment and avoidable deaths. Adding AI to the diagnosis process can limit human errors and improve health outcomes for patients.
Early in the COVID19- pandemic, KAUST developed a rapid and accurate method for COVID19- detection. Researchers developed an AI system that processed CT scans of patients. The image processing took less than one minute and correctly detected COVID-19 100% of the time.
KAUST researchers have recently advanced their AI models by limiting scanner-specific noise to increase CT scan image resolution. They also invented an AI algorithm to segment blood vessels and the trachea to a very fine resolution. These innovations are part of a pipeline that will lead to automatic, accurate and rapid pneumonia diagnosis.