A proposed system for gender classification using lower part of face image

Present study proposes a fast gender classification system from frontal facial images using features selected from mouth and chin only. In most of the study on gender classification found in literature deals with lots of features which makes the classification system a complex one whereas reducing the number of features makes the system simpler but selection of Present study proposes a fast gender classification system from frontal facial images using features selected from mouth and chin only. In most of the study on gender classification found in literature deals with lots of features which makes the classification system a complex one whereas reducing the number of features makes the system simpler but selection of features also plays important role in gender classification. Generally lower part of face image carries sufficient information regarding gender of a person. So in this study, features from lower part of face are considered for gender identification. Proposed method works in four steps-a) Extraction of the Lower part of frontal face images using the method geometric model proposed by Bhattacharjee et al. b) Construction of Gray Level Co-occurrence Matrix from the extracted image c) Extraction of Features from GLCM and d) Classification of the face using a standard classifiers. The proposed method has been tested on 75 male and 35 female color face images of standard FRAV2D database and some face images captured using standard camera. Experimental result shows the effectiveness of this simple gender classification system which achieves 94.34±1.8% accuracy on test data.features also plays important role in gender classification. Generally lower part of face image carries sufficient information regarding gender of a person. So in this study, features from lower part of face are considered for gender identification. Proposed method works in four steps-a) Extraction of the Lower part of frontal face images using the method geometric model proposed by Bhattacharjee et al. b) Construction of Gray Level Co-occurrence Matrix from the extracted image c) Extraction of Features from GLCM and d) Classification of the face using a standard classifiers. The proposed method has been tested on 75 male and 35 female color face images of standard FRAV2D database and some face images captured using standard camera. Experimental result shows the effectiveness of this simple gender classification system which achieves 94.34±1.8% accuracy on test data.