基于改进MeanShift算法的视线跟踪技术

Eye Tracking Technology Based on Improved Mean-Shift Algorithm

摘要:

介绍了SURF算法和MeanShift算法的基本原理。将MeanShift跟踪算法应用于视觉鼠标技术,提出了基于OpenCV技术的视线系统设计方法；提出了基于SURF算法的自动嘴部区域初始化MeanShift改进算法。实验结果表明,算法匹配识别效果较好,基于MeanShift跟踪算法的嘴唇跟踪系统实时性好、具有较好鲁棒性。

Abstract:

The basic concept of SURF and MeanShift algorithm was introduced. The application of MeanShift algorithm to eye tracking was made and the design method of eye tracking system was elaborated. The improvement of MeanShift algorithm about automatic mouth area initial depend on SURF algorithm was elaborated. Experiment result expressed that the algorithm was effective in matching and identifying, and lip tracking system depending on MeanShift tracking algorithm was good in real time and robust feature.