基于Hough变换和无轨迹卡尔曼滤波的眼睛角点跟踪

提出了一种基于改进Hough变换(HT)和无轨迹卡尔曼滤波(UKF)的眼睛外角点跟踪算法。该算法在输入图像中存在虹膜时采用改进Hough变换提取眼睑轮廓并得到眼睛外角点位置,当输入图像中检测不到虹膜时,采用UKF算法对当前帧眼睛角点进行估计。实验证明,本文算法能精确地跟踪眼睛外角点。

Eye corner tracking based on Hough transform and unscented Kalman filter

Eye features include iris, eyelids, eye corners etc. The tracking of eye features plays an important role in face recognition system as the eye features are among the most salient facial features. A robust algorithm for tracking the eye outer corners in video sequence was presented. This algorithm is based on modified Hough Transform (HT) and Unscented Kalman Filter (UKF). The proposed algorithm uses the modified HT to extract the eye outer corners when the iris is available in the input images. Otherwise the algorithm uses the UKF to estimate the positions of the eye outer corners. Experiments demonstrate the accuracy of the proposed algorithm.