Hangiandreou N J, Fetterly K A, Bernatz S N, Cesar L J, Groth D S, Felmlee J P
Department of Diagnostic Radiology, Mayo Clinic and Foundation, Rochester, MN, USA.
J Digit Imaging. 1998 Aug;11(3 Suppl 1):180-6. doi: 10.1007/BF03168299.
This study indicates that contrast-detail data should be very helpful in providing quantitative measurements of overall electronic display quality. The method would be suitable for new equipment selection, acceptance testing, and quality control. The recommended protocol would only involve observer data obtained using test images with mid-range background pixed values. Improvements to the current linear curve fit may also provide increased levels of measurement precision and sensitivity. To put the measurements in proper context, MTC measurements of a group of displays currently in use and deemed acceptable for the clinical display) should be obtained by a group of observers, if possible. When making quantitative recommendations regarding equipment selection, or display configuration (eg, maximum display luminance or ambient room lighting levels), a group of observers should be used, since the decisions made will presumably affect a large number of radiologists, technologists or clinical physicians using the display workstations. With a group of five observers, and using the group paired difference analysis technique, measurement precision will be 9.0%, and sensitivity to MTC changes will be 11.1%. Each set of raw data for a measurement of MTC can be collected and analyzed for each observer in approximately 30 minutes, so data sufficient for a comparison of two devices could be collected and analyzed within an hour. When making measurements for equipment acceptance testing or routine QC measurements (eg, on a quarterly or twice-yearly basis), measurements from a single observer should suffice since the goal is an assessment of the relative performance of an individual device. Precision of the single observer MTC measurements will be 6.8%, and sensitivity will be 15.2%. Measurements made over a period of time should have a reproducibility of about 5%.
本研究表明，对比度 - 细节数据对于提供电子显示器整体质量的定量测量非常有帮助。该方法适用于新设备选型、验收测试和质量控制。推荐的方案仅涉及使用具有中等范围背景像素值的测试图像获得的观察者数据。对当前线性曲线拟合的改进也可能提高测量精度和灵敏度。为了将测量结果置于适当的背景下，如果可能的话，一组观察者应获取一组当前正在使用且被认为适用于临床显示的显示器的调制传递函数（MTC）测量值。在就设备选型或显示配置（例如，最大显示亮度或室内环境光照水平）提出定量建议时，应使用一组观察者，因为所做的决策可能会影响大量使用显示工作站的放射科医生、技术人员或临床医生。对于一组五名观察者，并使用组配对差异分析技术，测量精度将为9.0%，对MTC变化的灵敏度将为11.1%。每次测量MTC的原始数据集，每位观察者大约可在30分钟内收集并分析完毕，因此在一小时内即可收集并分析足以比较两台设备的数据。在进行设备验收测试或常规质量控制测量（例如，按季度或半年一次）时，单个观察者的测量结果就足够了，因为目标是评估单个设备的相对性能。单个观察者MTC测量的精度将为6.8%，灵敏度将为15.2%。在一段时间内进行的测量应具有约5%的再现性。
