中文摘要：

针对无监督局部保持投影算法在头部姿态估计上的高误差性和对噪音的敏感性的问题,提出一种鲁棒的局部保持投影算法。其基本出发点是先对训练的头部姿态加以姿态标注,并获得各个头部姿态间的偏置距离,再对所有头部姿态样本点进行异常值的度量,训练出较好的线性映射矩阵。实验结果表明,改进的方法比传统的LPP在头部姿态估计上取得较好的效果。

英文摘要：

Aimed at the problems of the high head pose estimation error and noise sensitivity for unsupervised LPP,an robust LPP algorithm is presented.By this algorithm,the head poses of the training samples are firstly labeled,and all the biased distances between the head poses are calculated.And then the outliers of the head samples are detected,so as to train a better linear mapping matrix.This head pose estimation experiments show that the improved LPP achieve the better results than the traditional LPP.