用于瞳孔序列图像分析的实时轮廓检测和参数提取

本文介绍一种医用图像分析系统．该系统连续获取瞳孔图像，根据对图像序列的某些先验知识自动确定门限．在此基础上进行快速边缘检测，并自动修补缺损的目标轮廓．在判定所得轮廓有效后计算目标面积．若数据无效则自动重新采样．序列图像观测结束后对间隔不均匀的数据进行内插校正．系统所提供的分析结果可用于临床和医学研究

Real time Edge Detection and Parameter Extraction for Analysis of Pupil Image Series

A medical image analysis system is described in this paper. The system continuously acquires pupil images, and extracts the edges using an automatic thresholding algorithm. Defective object outlines are repaired and then validated before their areas are calculated. These are based upon certain a priori knowledge of the image series. If an invalid image is detected, a supplementary acquisition is carried out. This results in nonuniform intervals between individual images, and therefore necessitates an interpolation procedure in order to correct the final data series. The data provided by the system is useful both in clinical practice and medical research.