Rationale and Objectives. The purpose of this study was to measure the infl
uence of display luminance on detection performance and visual search behav
ior. The results of the study should be helpful in establishing minimally a
cceptable display conditions for viewing radiographs on cathode-ray tube (C
RT) monitors.
Materials and Methods. Two groups of six radiologists each viewed 50 pairs
of mammograms. One group viewed film images on a standard mammographic view
box; the other viewed images on a high-resolution CRT monitor. Two luminan
ce levels were studied for each display type. Observers reported on the pre
sence or absence of masses or microcalcification clusters and on their conf
idence in that decision. Confidence data were analyzed by using alternative
free-response receiver operating characteristic (AFROC) techniques. Eye po
sition also was recorded as observers viewed the images.
Results. For both the film and monitor studies, detection performance (AFRO
C area under the curve) was not affected significantly by display luminance
, but search behavior was. Total viewing and decision dwell times were shor
ter with the higher-luminance displays, especially for true-negative decisi
ons. Significantly more fixation clusters were generated during the search
of lesion-free than of lesion-containing images with the lower-luminance di
splays.
Conclusion. Display luminance affects visual search performance with both f
ilm and monitor displays without affecting detection performance significan
tly. Higher-luminance displays yield more efficient search performance. The
true-negative dwell times and number of clusters are suggestive that lower
-luminance levels prolong the search and recognition of normal, lesion-free
areas compared th lesion-containing areas.