基于半监督流形学习的人脸识别方法

Face Recognition Based on Semi-supervised Manifold Learning

如何有效地将流形学习(Manifold learning,ML)和半监督学习(Semi-supervised learning,SSL)方法进行结合是近年来模式识别和机器学习领域研究的热点问题。提出一种基于半监督流形学习(Semi-supervised manifold learning,SSML)的人脸识别方法,它在部分有标签信息的人脸数据的情况下,通过利用人脸数据本身的非线性流形结构信息和部分标签信息来调整点与点之间的距离形成距离矩阵,而后基于被调整的距离矩阵进行线性近邻重建来实现维数约简,提取低维鉴别特征用于人脸识别。基于公开的人脸数据库上的实验结果表明,该方法能有效地提高人脸识别的性能。

Recently,manifold learning and semi-supervised learning are two hot topics in the field of machine learning. However, there are only a few researches on how to incorporate semi-supervised learning and manifold learning, especially for face recognition. A new semi-supervised manifold learning for face recognition was proposed. This method relies on the distance matrix formed by both labeled and unlabeled samples, and then the local linear embedding （LLE） method was used to extract discriminative manifold features according to the modified distance matrix.The proposed method produces better classification performance which captures the intrinsic manifold structure collectively revealed by labeled and unlabeled samples. Experimental results on public face databases show that the proposed method can improves face classification performance effectively.