基于Mean Shift算法和粒子滤波器的人眼跟踪

基于视觉的驾驶疲劳检测是人脸表情识别技术最有商业前途的应用之一,实时人眼跟踪是其中的关键部分。为了解决跟踪方法对眼睛的部分遮挡、人脸尺度变化等过于敏感的问题,提出了一种综合MeanShift算法和粒子滤波器的跟踪算法。利用粒子滤波器得到样本的观测值后,将MeanShift分析用于每一个粒子,使得粒子集中在测量模型的局部区域内,很好地克服了粒子滤波器的退化现象并有效缩短了计算时间。实验结果表明该算法实时性强,且具有良好的鲁棒性。

Eye Tracking Based on Mean Shift Algorithm and Particle Filter

The vision-based driver fatigue detection is one of the most prospective commercial applications of facial expression recognition technology.Driver's fatigue is one of the chief causes of traffic accident.So it's very important to detect driver's fatigue status and decrease accident rate.And real-time eye tracking is the crucial part of it.But one common problem to eye tracking methods proposed so far is their sensitivity to lighting condition change,partial occlusion of eye,significant clutter,face scale variations and head rotations in depth.In this paper,to solve this problem,we present a tracking algorithm combining Mean Shift algorithm and particle filtering.After each sample is measured by observation,mean shift analysis is applied to each sample based on observation density.After Mean Shift iterations,samples are "herded" to the local modes of the observation.So the degeneracy problem is efficiently overcome and the computational cost is decreased.Experimental results show that this algorithm is robust and could track eyes satisfactorily.