The framework is called ItpCtrl-AI, which stands for interpretable and controllable artificial intelligence, and is “transparent and highly accurate,” according to a Feb. 25 news release from Fayetteville-based University of Arkansas.
A heat map was created from recordings of radiologists’ eye movements as they reviewed chest X-rays. The heat map was then used to teach the AI tool where to search for abnormalities, the release said.
The researchers are now working with Houston-based University of Texas MD Anderson Cancer Center to train ItpCtrl-AI on CT scan interpretation.
Read the full study here.
