Face Recognition Using Improved-LDA
This paper introduces an improved-LDA (I-LDA) approach to face recognition, which can effectively deal with the two problems encountered in LDA-based face recognition approaches: 1) the degenerated generalization ability caused by the “small sample size” problem, and 2) Fisher criterion is nonoptimal with respect to classification rate. In particular, the I-LDA approach can also improve the classification rate of one or several appointed classes by using a suitable weighted scheme. The key to this approach is to use the direct-LDA techniques for dimension reduction and meanwhile utilize a modified Fisher criterion that it is more closely related to classification error. Comparative experiments on ORL face database verify the effectiveness of the proposed method.
KeywordsFace Recognition Linear Discriminant Analysis Face Image Scatter Matrix Fisher Criterion
Unable to display preview. Download preview PDF.