摘要
针对人脸面部特征提取技术是自动视觉翻译和人脸识别中的关键技术之一,介绍了基于统计的人脸形状模型建立原理。利用有标志点的图像训练集变化模式的先验知识而建立模型的方法能够将人脸形状的变化限制在合理的范围内,并能够充分表征人脸特征,实验证明该方法是有效的。
Facial features extraction is one of the essential techniques in automated visual interpretation and recognition of human face. The paper introduces the method for building face models by learning patterns of variability from training set of correctly annotated images. The face models can only deform in ways found in the training set,so it can interpret facial features sufficiently. The result of the experiment shows that the method is effective.
出处
《河北工业科技》
CAS
2006年第1期12-14,共3页
Hebei Journal of Industrial Science and Technology
关键词
标定特征点
图像配准
主成份分析
labeling the training set,aligning the training set
PCA
