首页> 外文会议>Videometrics, range imaging, and applications X >Analysis of keratoscopic images for detecting fixational eye movements and ocular surface deformation
【24h】

Analysis of keratoscopic images for detecting fixational eye movements and ocular surface deformation

机译:分析用于检测注视眼动和眼表变形的角膜镜图像

获取原文
获取原文并翻译 | 示例

摘要

A sequence of videokeratoscopic images was registered using commercially available instrument E300 at a rate of 50rnfps. During the 20 seconds measurement, subject's head was fixed strongly. Acquired images were analyzed forrndetecting fixational eye movements and corneal surface deformation. For this purpose two rings were extracted fromrneach frame and the ellipses were fitted to them, using least square method. The time series of the ellipses geometricalrnparameters were considered: minor and major axes length as well as the ellipses center and the orientation. Thernfrequency spectra of mentioned parameters were obtained by application of the Fast Fourier Transform. The longitudinalrnposition of the corneal apex was controlled, thanks to the cone side viewer installed inside the videokeratoscope.rnThe average amplitude of the variation of the ellipse's axes length is around 20μm and of the orientation of the ellipsernaround 0,1 rad. In the signals frequency characteristics, appear the peak corresponding to the heart rate. No clearrnrelationship was found between the variations of the fitted ellipse parameters and the longitudinal position of the cornealrnapex.rnThe fixational eye movements were examined using two different methods. One of them consists of calculating therncorrelation function between the first and successive frame of the sequence and searching its maximum. The other isrnbased on tracking the center of the ellipse fitted to particular ring of the videokeratoscopic image. The accuracy of thernsecond method found to be higher.rnSimple methods proposed in this work can extend the application of videokeratoscopic measurements.
机译:使用市售仪器E300以50rnfps的速率记录一系列视频角膜镜图像。在20秒的测量过程中,受试者的头部牢固固定。分析获取的图像以检测注视眼的移动和角膜表面变形。为此,从每个框架中提取两个环,并使用最小二乘法将椭圆拟合到它们。考虑了椭圆几何参数的时间序列:短轴和长轴的长度以及椭圆的中心和方向。通过快速傅里叶变换获得上述参数的频谱。得益于安装在视频角膜镜内的锥形侧面观察器,角膜顶点的纵向位置得到了控制。椭圆轴长度变化的平均幅度约为20μm,椭圆的取向约为0.1 rad。在信号频率特性中,出现与心率相对应的峰值。在拟合的椭圆参数的变化与角膜顶点的纵向位置之间没有发现明确的关系。使用两种不同的方法检查了固定眼的运动。其中之一包括计算序列的第一帧和后续帧之间的相关函数并搜索其最大值。另一个基于跟踪适合视频角膜镜图像的特定环的椭圆的中心。发现第二种方法的准确性更高。这项工作中提出的简单方法可以扩展视频角膜镜测量的应用。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号