基于眼底OCT图像识别青光眼病症的算法研究

青光眼是目前常见的眼部疾病之一,具有较强的隐蔽性和突发性。由于光学相干层析成像(OCT)具有无创性、高探测灵敏度等特点,在临床上被广泛用于眼科疾病的诊断,但如何准确、高效地基于OCT眼底图像识别青光眼病症仍然是一个难题。为此提出了一种新的基于眼底OCT图像识别青光眼病症的算法,该算法首先对OCT图像进行直方图均匀化,解决其灰度不均匀的问题,然后再用数学形态学闭操作去除图像中视盘周围血管与神经干扰,之后使用自适应阈值分割和边缘检测canny算法对视盘视杯进行边界追踪,最后统计特征区域内像素个数获取CDR参数。通过实验将本方法应用于成人OCT眼底图上,结果显示本方法提取出的CDR参数与专家手绘CDR参数相比平均相似率在94. 9%以上,可以作为临床诊断的定量分析工具。

Research on algorithms for glaucoma recognition based on fundus OCT image

Glaucoma is one of the most common eye diseases with strong concealment and sudden onset.Optical coherence tomography(OCT)is widely used in the diagnosis of ophthalmic diseases because of its non-invasive and high detection sensitivity.However,it still remains a difficulty that how to recognize glaucoma diseases accurately and efficiently based on fundus OCT images.In this paper,a new algorithm for glaucoma recognition based on OCT image is proposed.First,OCT image is homogenized by histogram to solve the problem of uneven gray scale.Then,the vascular and nerve interference around the optic disc is removed by mathematical morphology closed operation,and then adaptive threshold segmentation and edge detection Canny algorithm are used to track the boundaries of the optic cup.Finally,the CDR parameters are obtained by counting the number of pixels in the feature area.This method is applied to adult OCT fundus maps through experiments.The results show that the average similarity rate of CDR parameters extracted by this method is more than 94.9%compared with the hand-drawn CDR parameters of experts.So it can be used as a quantitative analysis tool for clinical diagnosis.