Classifier Independent Subbands Selection based on Information-Theoretic
Affan Alim, Imran Naseem · 2018
In this research, a consolidated framework is proposed for the automatic selection of the most discriminant subbands for the problem of face recognition. The subbnds selection method is independent of classifier. Essentially, a Local Binary Pattern (LBP) is used for transforming the face image into texturized face image undergo to wavelet packet decomposition which produces the several detail and approximation images. An information-theoretic method is used to select the most important subbands which are independent of classifiers. Two classifiers chi-square and support vector machine are used to verify that selected subbands are independent of classifiers. The proposed algorithm are evaluated on several databases and are shown to pick the most significant subbands which produce the better performance.