基于裂隙灯显微图像的眼前节特征提取的新方法

裂隙灯显微图像是包含最直接和最丰富眼前节信息的无创数据源，针对其获取和处理等方面的特殊性，需要一种能够较好地分析和利用裂隙灯显微图像的方法。本研究提出了一种对直接用CCD从裂隙灯显微镜采集到的图像进行处理和分析的方法。首先采用一种改进的Canny算法进行边界提取，其中应用了多尺度的算法计算图像梯度。进而对所获得的边界点进行分类识别和曲线拟合等处理，获得眼前节各组织结构的边界，最后据此提取出眼前节的各项特征参数。实验获得了较好的结果，对临床有一定的指导意义。本方法具有操作简便，运算速度快，主观影响小和适应性强等特点。

A NEW METHOD BASED ON SLIT LIGHT MICROGRAPH TO OBTAIN ANTERIOR EYE SEGMENT CHARACTER

Slit light image is non invasive data source containing the most direct and abundant information of the anterior eye segment. Because of the particularity of the obtainment and process, a method for analyzing and making use of the slit light image is required. In this paper, a method of processing and analyzing the image obtained by CCD from slit light was proposed. First, boundaries were extracted by an ameliorated canny arithmetic. A multi measure arithmetic was used to calculate the grads of the image. Then the points obtained were identified apart and the curves were fitted according to them. The boundaries of each tissue structures of anterior eye segment were acquired. The characteristic parameters of anterior segment were picked up based on above results. This method was proved easy to be operated, fast, less subjective and more adaptive.