Age Detection based on Facial Features Using Support Vector Machine

Abstract

Face aging has been studied for decades. Determining age from a facial shot is key to our technique for diagnosing abnormal behavior. Security monitoring, forensics, biometrics, and Human-Computer Interface (HCI) use facial age estimates. We only look at adults 1–75 in the UTK Face database, which covers 0 to 116 years. The database contains 23,708 face photos with age, gender, and ethnicity annotations. In work, preprocessing, feature extraction, feature selection, and age categorization are involved. Preprocessing adjusts images. Computer vision uses Local Binary Pattern (LBP) and Histogram of Oriented Gradient (HOG) as visual descriptors, whereas fscmrmr is utilized for classification. Support Vector Machines (SVM) improve classification accuracy in highdimensional areas. The chosen characteristics are concatenated and passed to a multiclass SVM classifier to classify the images with 95.69% success.