Ward N P, Tomlinson S, Taylor C J
Department of Medical Biophysics, University of Manchester, England.
Ophthalmology. 1989 Jan;96(1):80-6.
A computer-based image analysis system was used to detect and measure exudates in fundus photographs. A fundus transparency was imaged, digitized, and stored in image memory. The stored image was then processed by several operators, to reduce shade variations in the image background and enhance the contrast between this background and the exudates. Exudates were separated from the background on the basis of their brightness or "gray level" and were then copied in to a binary image. For comparative purposes, the binary image was superimposed on the original unprocessed image. Exudate areas were measured using the binary image, which was also transferred to a printer to provide a permanent record or "exudate map." The system was able to discriminate between standard photographs used to grade hard exudates in the Early Treatment for Diabetic Retinopathy Study (ETDRS). It was also used to monitor the response of a subject to laser treatment.
一个基于计算机的图像分析系统被用于检测和测量眼底照片中的渗出物。对眼底透明度进行成像、数字化处理并存储在图像存储器中。然后，存储的图像由多个算子进行处理，以减少图像背景中的阴影变化，并增强该背景与渗出物之间的对比度。渗出物根据其亮度或“灰度级”与背景分离，然后被复制到一个二值图像中。为了进行比较，将二值图像叠加在原始未处理图像上。使用二值图像测量渗出物面积，该二值图像也被传输到打印机以提供永久记录或“渗出物图谱”。该系统能够区分糖尿病视网膜病变早期治疗研究（ETDRS）中用于对硬性渗出物进行分级的标准照片。它还被用于监测受试者对激光治疗的反应。
