Gender classification based on similarity features through SURF and SVM

Abstract

The recognisable proof of people in view of their biometric body parts, for example, face, fingerprint, walk, iris, and voice, assumes an imperative part in electronic applications and has turned into a prominent territory of research in image pre-processing. It is likewise a standout amongst the best utilisations of computer-human interaction and understanding. Out of all the previously mentioned body parts, the face is one of most well known qualities in view of its extraordinary feature. In reality, people can process a face in an assortment of approaches to characterise it by its personality, alongside various different attributes. In this paper, we proposed a new algorithm to extract the facial features using SURF algorithm, features are invariant to extract affine transformations are extracted from each face using speeded up robust features (SURF) method (Morteza and Yousefi, 2011) and shows best accuracy on real-time face images compared with different licence datasets like ORL database and FGNet database and with different training ratios by using SVM algorithm (Rahman et al., 2013; Moghaddam and Yang; 2000; Swaminathan, 2000).