Detection of Low-Contrast Signals

Publisher Website
Google Scholar
Abstract
Sixteen observers averaging 7.4 years experience in interpreting radiology images and 16 observers withoout radiology training were tested for their ability to detect low-contrast signals in structured noise backgrounds using a digital display system. No significant diffrences in true positives, false positives, time to complete test, or a calculated "performance value" could be detected between the two groups. Data from the beginning and end of the test indicated that no difference in performance between groups evolved as the test progressed. Therefore, radiology training seems to offer no significant advantage in low-contrast signal detection in an unfamiliar structured noise background, nor in rapid development of better search strategies of decision criteria.