基于预处理的Ballon模型及在眼疲劳检测中的应用

获取眼睛轮廓和眼球轮廓对进一步研究人眼疲劳状态具有重要意义.提出基于预处理的Ballon模型对含有双眼的局部人脸图片进行人眼和眼球检测的方法.用微分算法抽取边缘、标识出物体、使用几何特性识别人艰区域,自动获取主动轮廓所需要的眼睛初始轮廓;使用提出的基于预处理的Ballon模型收敛人眼;结合数学形态等方法收敛到眼球.实验结果表明,该方法具有效果好、鲁棒性好、受光线影响弱等优点.

Ballon snake model based on pretreatment and its application in tiredness drive

To obtain contour of eyes and eyeballs is great meaningful for the future research about closed eyes. A method to detect them using Ballon snake model based on pretreatment in a picture with human eyes is presented. Firstly, we can detect edge in differential algorithm and mark objects and find out location of eyes according to geometric properties and construct eyes area, and the initial contour of eyes is obtained automatically. Then, we can converge contour of eyes using Ballon snake model based on pretreatment. Finally, combined with mathematical morphology, the contour of eyeballs is obtained using snake model again. Experiments showed that the method has advantages of robust and small impact on light effect.