Automatic Human Face Texture Analysis for Age and Gender Recognition
Keywords:
Age Estimation, Gender, Estimation, Facial, Recognition, HOG, KNNAbstract
Facial recognition and textural analysis have been a hot research area for decades. Facial variation, mainly caused due to the factor of age, change in expression and gender, are interesting and evolving research areas from the domain of Facial recognition. Human face is a toolkit to work these variations and explore the accurate gender, age and gesture of the person which can be used to assist in many vision applications. Major focus of this research is to design an age and gender estimation algorithm to increase estimation rate. My proposed system uses HOG binary descriptors for features extraction. KNN classifies between the features from different classes of age and gender. The evaluation is performed on two publicly available databases i.e. FG-NET and IMM databases for estimation of age and gender respectively. Proposed method (HOG + KNN) achieves estimation rate of 89.7% for age detection on FG-NET subset database and 56% over entire database. Similarly, accuracies of gender recognition are as high as 94.16%