【摘要】 研究了复杂背景下多姿态人脸图像中的人眼检测方法.首先,将得到的图像映射到YCbCr空间,利用肤色在YCbCr颜色空间的分布特性建立肤色模型,分割出肤色并得到人脸的区域;再次,利用人眼的分布特点及自身特性预先设定一系列的规则,通过规则得到可能的人眼区域;最后,设计了人眼模板分布模型,并利用该模型对人眼位置进行最后的确认.实验表明,该方法具有较高的检测速度和检测正确率.更多还原

【Abstract】 Eye detection method for multi-pose face image in complex background is studied.Firstly,we project the image to YCbCr space and get skin and face region by the skin model based on the distribution of the skin;secondly,the possible eye regions can be found according to the rules formulated by the eye distribution and its characteristics;lastly,the eye position is confirmed by eye templates distribution model.Experiment results show that the method is efficient and precise.更多还原