论文部分内容阅读
主分量分析(PCA)和线性鉴别分析(LDA)是模式识别领域使用广泛的两种特征抽取方法.本文针对两种方法的不足之处,并从样本分布相似度出发提出一种期望分布鉴别分析(EDDA)方法,抽取到的鉴别特征的总体分布和设定的期望分布最为相近.即通过 EDDA 得到的投影向量可以抽取出最接近理想分布的鉴别特征.EDDA 在投影向量的求解问题上不存在小样本问题,抽取的鉴别特征维数小,并且整体识别性能得到增强.在 ORL、Yale 人脸库上的实验结果证明本文方法在人脸识别精度上优于 PCA 和 LDA 方法.
Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA) are two widely used feature extraction methods in the field of pattern recognition.This paper aims at the shortcomings of the two methods and proposes an Expected Distribution Discriminant Analysis (EDDA) method, the overall distribution of the extracted discriminant features is most similar to the expected expected distribution, ie, the projection vector obtained through EDDA can extract the discriminant features that are closest to the ideal distribution.The EDDA does not exist in the solution of the projection vector The small sample size and the small number of discriminant features are extracted, and the overall recognition performance is enhanced.Experimental results on ORL and Yale face database demonstrate that our method is superior to PCA and LDA in face recognition accuracy.