Study On Face Classification Methods Based Facial Features

Face recognition, which has a prospective application foreground, is one of the most active and challenging problems in pattern recognition and machine vision field. Face classification based on facial features is one example which can be applied to many situations.This paper studied frontal faces, and classified the facial images, based on two attributes——sunglasses and masks, into four categories: faces, faces with sunglasses, faces with both sunglasses and masks, faces with masks.Generally, face classification system has the following three steps: face detection, feature extraction, and classification. In this paper, we mainly studied feature extraction and classification algorithms, including Principal Component Analysis (PCA), Linear Discriminant Analysis (LDA), Correlation Coefficient (Correlation), Support Vector Machine (SVM), and AdaBoost. Based on previous works, we researched on these current algorithms and made experiments for five classification methods: PCA+LDA+Correlation,SVM,PCA+LDA+SVM,SVM+AdaBoost,and PCA+LDA+SVM+AdaBoost.The experimental result showed that using PCA+LDA to reduce feature dimensions could decrease the classification time largely, with very little harm to classification performance. In addition, a classification approach, a combination of SVM and AdaBoost, which exploits the strong structure of faces to select and train on the optimal set of features for each attribute was proposed in this paper, and the result seemed to be ideal.