A Novel Approach for Blurred Face Recognition System Using GLDA Features with LCDR Classification

摘要

Objective: To design an effective face recognition system invariant to all image degradation parameters. Methods: The proposed system designed an efficient image restoration based on Iterative Graph based Image Restoration technique. It provides high reconstruction rate. Gabor Linear Discriminant Analysis (GLDA) feature extraction method is used to extract features for the restored faces images and Linear Collaborative Discriminant Regression Classifier (LCDRC) is adopted. GLDA based feature extraction is the combination of both the features like Gabor and LDA which are used to obtain the maximum recognition rate. Findings: The LCDRC gives a discriminant subspace with maximum collaborative between-class reconstruction error and minimum within-class reconstruction error. Applications/Improvements: It is an improvement over the linear discriminant regression classifier (LDRC). From the experimentation results, it has been achieved a recognition rate of 99.2% even in the case of blurred face images.