OPTIMIZING SOFTCOPY DISPLAYS FOR DIGITAL MAMMOGRAPHY

乳腺癌是当今美国女性最常见的确诊癌症，也是导致癌症死亡的第二大原因。尽管在癌症治疗和预防方面取得了许多进展，但早期发现乳腺癌仍然是有效治疗和降低死亡率的最佳机会。多年来，传统的X射线乳房X光检查一直是乳房成像的黄金标准，并已被证明可以降低乳腺癌的死亡率。然而，使用传统的乳房X光检查仍然有大约30%的病变漏检率，延误了相当数量的女性的发现和治疗。数字化乳房X光检查为显著提高诊断效率和准确性提供了潜力，但与任何新的成像技术一样，需要大量工作来优化成像链中各个组件的设计和操作参数。数字乳房X光摄影链中一个极其重要的环节是显示器。虽然在地平线上有许多有趣的显示介质(例如，液晶显示器、场发射显示器)，但最常见、广泛使用和廉价的显示器是阴极射线管(CRT)监视器。然而，在优化CRT显示器以显示数字乳房X光照片方面，仍有许多工作要做。该计划的目标是开发一种有效的方法来优化数字乳房X光摄影系统的显示性能。我们将讨论PA(PA-99-082)“数字乳腺摄影显示器和工作站的发展”的三个方面：(1)预测模型，(2)感知和(3)工作站(显示器)设计。特别是，放射科医生的表现将是衡量几个显示特征的标准。然后，这些结果将用于增强和验证人类视觉性能模型(JNDmetrix Vision Model)。然后，JND Metrix模型将用于系统地探索和优化当前和未来数字乳房X光显示系统的CRT设计权衡。

Breast cancer is the most commonly diagnosed cancer and the second leading cause of cancer deaths among American women today. Despite many advances in cancer treatment and prevention, early detection of breast cancer remains the best opportunity for effective treatment and low mortality. Conventional x-ray mammography has been the gold standard for breast imaging for many years, and has been shown to reduce mortality due to breast cancer. However, there remains about a 30 percent lesion miss rate using conventional mammography, delaying detection and treatment for a significant number of women. Digital mammography offers the potential for significant improvements in diagnostic efficiency and accuracy, but as with any new imaging technology, a great deal of work is needed to optimize design and operating parameters in various components of the imaging chain. One extremely important link in the digital mammography chain is the display. Although there are a number of interesting display media on the horizon (e.g., liquid crystal displays, field emission displays), the most common, widely available and inexpensive display is the cathode ray tube (CRT) monitor. However, there is still much work to be done to optimize the CRT monitor for display of digital mammograms. This program's goal is to develop an efficient method of optimizing display performance for digital mammography systems. We will address three aspects of the "Development of Digital Mammography Displays and Workstations" PA (PA-99-082): (1) predictive models, (2) perception and (3) workstation (display) design. In particular, radiologist performance will be measures for several display characteristics. These results will then be used to enhance and validate a model of human visual performance (JNDmetrix Vision Model). The JND metrix model will then be used to systematically explore and optimize CRT design tradeoffs for current and future digital-mammography display systems.