Kundel H L, Mezrich J L, Brickman I, Siegel R, Miller W T, Epstein D M, Gefter W B, Arenson R L, Seshadri S B, Khalsa S
Department of Radiology, University of Pennsylvania, Philadelphia 19104.
Radiology. 1987 Dec;165(3):747-52. doi: 10.1148/radiology.165.3.3685354.
A classification test was used to compare the results of digitizing chest radiographs with a linear diode array camera and a laser scanner. The ability of B readers to classify a pneumoconiosis test set from original radiographs and the digitized images was compared. The classification of profusion was impaired when images were digitized with the diode array camera but not when they were digitized with the laser scanner. The impairment found with images digitized with the linear diode array camera was overcome with use of a zoom function at the display station. The size of areas of opacification was overestimated on images from both scanners. Variation between and within observers (kappa approximately equal to 0.6) was consistent over observers and over modalities.
采用分类测试来比较使用线性二极管阵列相机和激光扫描仪对胸部X光片进行数字化处理的结果。比较了B级阅片者对尘肺病测试集的原始X光片和数字化图像进行分类的能力。当使用二极管阵列相机对图像进行数字化处理时,小阴影密集度的分类受到损害,但使用激光扫描仪进行数字化处理时则未受影响。通过在显示工作站使用缩放功能,克服了线性二极管阵列相机数字化图像时出现的损害。两种扫描仪生成的图像上,阴影区域的大小均被高估。观察者之间和观察者内部的差异(kappa约等于0.6)在不同观察者和不同模式下是一致的。