Face recognition using PCA with GNP-fuzzy data mining

Conventional Principal Component Analysis (PCA) based face recognition methods measure the similarity between two feature vectors by some simple geometric distances, such as Euclidean distance and cosine distance. It largely limits the correct recognition rate of such methods in different conditions of appearance, light and pose. Thus, this paper proposes a novel approach for face recognition using PCA with Genetic Network Programming (GNP) fuzzy (GNP-fuzzy) data mining. Different from conventional methods, the proposed method uses the class association rules to represent the features of a face and calculate the similarity between two feature vectors of two different faces. Experimental results have also demonstrated the effectiveness of the proposed method in complex test environments.