通过人脸分析方法估计人类年龄的困难在于人脸外观的变化原因除了年龄变化,还受生活方式及环境等影响。人脸图像在采集时的复杂性造成的光照不均,人脸姿势等,也增加年龄估计难度。目前大多数年龄估计的方法都是预先对人脸图像进行灰度均衡和人脸矫正等预处理,采用外形或纹理信息作为特性的估计方法。提出一种多特征融合的人脸年龄估计方法,采用有较好的光照及旋转不变性的局部二进制模式(LBP)和梯度直方图(HOG)作为人脸年龄变化的特征描述子,用典型相关分析法(CCA)在特征层将LBP和 HOG融合成更具年龄变化鉴别力的特征。然后通过学习得到一个多线性回归函数揭示融合后的特征和年龄之间的关系。实验结果表明该方法在没有人脸矫正等预处理的情况能取得较好效果。
Estimating human age via facial image analysis is very difficult ,due to the fact that the factors of causing variations in the appearance of the human face include not only the aging ,but also the lifestyle and life environments etc .Both illumination and position of facial image have side-effect on the age estimation . Existing estimation methods consider the shape or texture of facial image to characterize human aging with the preprocessing of the gray-balance and Procrustes analysis .Motivated by the fact that both LBP and HOG information of facial images are robust to control illumination and rotation and can provide complementary information in characterizing human age ,we propose fusing these two sources of information at the feature level by using canonical correlation analysis (CCA) for enhanced facial age estimation .Then , we learn a multiple linear regression function to uncover the relation of the fused features and the ground-truth age values for age prediction .Experimental results are presented to demonstrate the efficacy of the proposed method .