GRAY-SCALE IMAGE PROCESSING FOR DIGITAL MAMMOGRAPHY

筛查乳房X光检查已被证明是一种有效的方法 识别早期乳腺癌。乳房X光检查发现的癌症倾向于 比乳房物理检查发现的更小且不太晚期的阶段 考试。较小和较低分期的乳腺癌存活率较高 费率。不幸的是，大约10%的乳腺癌是看不见的 进行乳房X光检查，特别是在有大量乳房的患者 腺体组织。如果我们能通过以下方式提高乳房X光检查的灵敏度 通过灰度级处理使记录的信息可感知 技术(即“对比度增强”或“图像处理”) 这些癌症更容易显现，及早发现可能会导致 显著降低了这一人群的死亡率。这样做的目的是 建议对传输数据的不同方法进行评估 记录在乳房X光照片中以显示强度，以便增加 提高乳房X光检查的敏感性，促进乳房的早期识别 癌症。 建议的研究包括三个部分：第一，我们将提供初步的 灰度处理方法，以增强标准显示的数字 临床现场的图像。这些最初的方法将来自于结果 在北卡罗来纳大学在预设强度窗口上进行的实验，以及 开发具有适当人体工程学的软拷贝显示器。第二，我们 将我们的图像显示处理方法应用于一系列有前景的 评价不同静态和动态影响的临床研究 图像处理方法在心脏异常检测中的应用 从两个临床站点(MGDM和MGDM)获得的数字化乳房X光照片 TJUDM)。改进的图像处理算法和软拷贝人体工程学将 在整个赠款期间按原样提供给临床站点 可用。第三，我们将进行受控观察者研究，以确定 无论是数字乳房X光片，使用有损压缩和解压缩 由GE远程放射摄影组提供的压缩算法，包含 与原始图像相同的临床信息。

Screening mammography has proven to be an effective procedure in identifying early breast cancer. The cancers found by mammography tend to be smaller and of less advanced stages than those found by breast physical examination. Smaller and lower stage breast cancers have better survival rates. Unfortunately, approximately 10% of breast cancers are not visible with mammography, particularly in patients with large amounts of breast glandular tissue. If we can increase the sensitivity of mammography by making perceivable the recorded information through greyscale manipulation techniques (i.e. "contrast-enhancement" or "image processing") so that these cancers are more readily apparent, early detection may result in significantly reduced mortality in this population. The aim of this proposal is to evaluate different methods of transferring the data recorded in a mammogram to displayed intensities, in order to increase the sensitivity of mammography and facilitate earlier identification of breast cancer. The proposed research has three parts: first, we will provide initial greyscale processing methods to augment a standard display of the digital images at the clinical sites. These initial methods will come from results of experiments conducted at UNC on preset intensity windows, and development of softcopy display with appropriate ergonomics. Second, we will apply our image display processing methods to a series of prospective clinical studies to evaluate the affect of various static and dynamic image processing methods on the detection of abnormalities in the digitally acquired mammograms from the two clinical sites (MGDM and TJUDM). Improved image processing algorithms and softcopy ergonomics will be provided to the clinical sites throughout the grant period as they are available. Third, we will conduct a controlled observer study to determine whether a digital mammogram, compressed and decompressed using a lossy compression algorithms provided by the GE telemammography group, contains the same clinical information as the original image.