A method for face recognition based on invariant eigenvectors and Hausdorff Fraction Distance is proposed. With this method, the invariant eigenvectors based on the image edge are firstly extracted. Then by computing the Hausdorff Fraction Distance between the invariant eigenvectors, the process for similarities evaluation is accomplished. Experimental results on the ORL face database validate that the proposed method is invariant to image rotation, minute edge alteration and illumination conditions, and can improve recognition precision and reduce time complexity simultaneously.
展开▼
