基于协方差矩阵表示的图像集匹配

Image Set Matching Based on Covariance Matrix

提出了一种图像集合匹配方法,该方法通过协方差矩阵对图像集合建模,利用对称正定的非奇异协方差矩阵构成黎曼流形上的子空间,将图像集的匹配转化为黎曼流形上的点的匹配问题.在ETH80和Honda UCSD数据库分别进行了基于图像集合的对象识别和人脸识别实验,分别达到96%和95.9%的识别率.

An image set matching method is proposed, in which the problem of the image set matching is for- mulated as matching points lying on the Riemannian manifold spanned by symmetric positive definite （SPD）, i.e. nonsingular covariance matrices. The similarity between two image sets is converted into the distance between two points in the Riemannian manifold. The proposed method is evaluated in set-based object classification and face recognition tasks, extensive experimental results show that the proposed method outperforms other state of the art set-based object matching and face recognition methods with recognition rate of 96% and 95.5% in the ETH80 ob- ject database and HondaUCSD video database, respectively.