Cyclic Exploration-Based Whale Optimization to Linear Discriminant Regression Classification for Face Recognition

Abstract

Generally, face recognition determines the judgment of whether or not a specific face is “known.” Moreover, the face identification denotes the retrieval of data or information about the “owner” of face. Under this concept, a number of researches are there in progression. Still, the research works are not yet up to the mark as human brain recognition. The intelligence in face recognition should be enhanced with high accuracy rate, and this approach tends to present a novel face recognition pattern with a concept of feature extraction and classification. The features are extracted using Active Appearance Model (AAM). Then the classification is done via linear collaborative Discriminant regression classification (LCDRC) model proposed by Xiaochao Qu. In the LCDRC classifier, the most important evaluation is the projection matrix that might get multiplied to the features while classification. The projection matrix must be optimal, so that the recognition accuracy can be greatly attained. In order to select the optimal projection matrix, this paper presents a Cyclic Exploration-based Whale Optimization model (CEWO), which is the modified form of Whale Optimization Algorithm (WOA). The comparison of the proposed face recognition model is done with the performance across the additional conventional techniques with regard to measures such as Accuracy, Precision, False positive rate (FPR), False negative rate (FNR) of the proposed model is proven.