Development of automatic gender classification algorithms for person specific identification

Abstract

Face is a very important biometric feature of humans. Automatic newlinerecognizing and analyzing of a face is one of the challenging tasks in object newlinerecognition. Classifying face images in terms of soft biometric qualities, such as newlinegender, age or ethnicity, has been getting an extensive amount of attention in the newlinemodern computer vision literature, especially in the video surveillance point of newlineview. In real-time, detection of faces and gender classification has become a newlinesignificant challenge due to the variations in head pose, scale variation in face newlineimages, different lighting conditions, partially occluded faces and noisy face newlineimages. So the robustness of the gender classification system has to be improved newlineby means of efficient image processing techniques. Human gender classification methods use iris, body shapes or gait cycle. However, the majority of studies used facial characteristics to distinguish gender. Automatic gender classification is an important area of research and it has great potential in the field of image processing. In recent years, this newlinemotivating area of research contributes to the development of machine learning, newlineimage processing, and human interaction domains. In the face recognition newlinesystem, if it identifies the gender first, then it will be easy to search a part of the newlinedatabase. So the searching time will be reduced and can achieve high newlinecomputational speed. newline newline