There is great interest in developing more advanced computer analysis and AI applications in radiology. As these methods advance, seamless human-computer interfaces are needed that can capture and integrate the interaction data, according to presenter Dr. Oguz Akin of Memorial Sloan Kettering Cancer Center in New York City.
Akin with fellow principal investigator Yingli Tian, PhD, from the City College of New York City, and colleagues developed a real-time system to track a radiologist's gaze and searching that could automatically capture -- in a setting identical to clinical practice -- attention heat maps and also spatial and temporal coordinates of the radiologist's search pattern while interpreting a study. The project is part of an ongoing collaboration between radiologists and computer scientists to develop AI technology in radiology, Akin said.
"This system can be a valuable tool to study the differences of attention and search among radiologists, to measure trainees' search patterns as compared to experts, to assess radiologist fatigue, or to better understand the reasons for missed lesions," Akin told AuntMinnie.com. "Another future application of this system could be the integration of these data with artificial intelligence to create a unique human-computer interface."
How did they do it? Check out this Wednesday afternoon talk to find out.
