Performance analysis of combination of CS with PCA and LDA for face recognition
Preeti, Dinesh Kumar · 2017 International Conference on Information, Communication, Instrumentation and Control (ICICIC) · 2017
Feature selection is one of the most vital processes to achieve maximum accuracy in the face recognition system which helps in selecting the important and relevant information. In this paper, Cuckoo Search algorithm is used for further selection of optimized set of features out of the features extracted using DCT (Discrete Cosine Transform) — PCA (Principal Component Analysis) (DCT-PCA+CS) and LDA (Linear Discriminate Analysis) (LDA+CS) methods. On comparison with LDA and PCA alone, it has been observed that the proposed work not only produces good recognition rate with the original face images but also with image even after these are motion blurred.