Face Recognition using the LCS algorithm

Today, the topic of human identification based on physical characteristics is a necessity in various fields. As a biometric system, a facial recognition system is fundamentally a pattern recognition system that identifies a person based on specific physiological or behavioral feature vectors. The feature vector is typically stored in a database upon extraction. The main objective of this research is to study and assess the effect of selecting the proper image attributes using the Cuckoo search algorithm. Thus, the selection of an optimal subset, given the large size of the feature vector dimensions to expedite the facial recognition algorithm is essential and substantial. Initially, by using the existing database, image characteristics are extracted and selected as a binary optimal subset of facial features using the Cuckoo algorithm. This subset of optimal features are evaluated by classifying nearest neighbor and neural networks. By calculating the accuracy of this classification, it is clear that the proposed method is of higher accuracy compared to previous methods in facial recognition based on the selection of significant features by the proposed algorithm.