JISE


  [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12]


Journal of Information Science and Engineering, Vol. 35 No. 1, pp. 105-123


Gender Classification with Jointing Multiple Models for Occlusion Images

CHIAO-WEN KAO1, HUI-HUI CHEN2,*, BOR-JIUNN HWANG2,
YU-JU HUANG1 AND KUO-CHIN FAN1
1Department of Computer Science and Information Engineering
National Central University
Taoyuan, 320 Taiwan
2Department of Computer Communication and Engineering
Ming Chuan University
Taoyuan, 333 Taiwan
E-mail: {chiaowenk; louise121307}@gmail.com;
{huichen; bjhwang}@mail.mcu.edu.tw; kcfan@csie.ncu.edu.tw

A facilitated and effective gender recognition approach is desirable for various applications such as for intelligent surveillance systems, human-computer interactions, and consumer behavior analysis. Since the human face conveys clear sexual dimorphism, the use of facial features seems an intuitive way to recognize gender. This paper proposes an efficient gender classification method using multiple classifiers to overcome the occlusion problem. The experiment is tested via 5-fold cross validation on the FERET and AR databases to evaluate the performance. The results show the proposed approach achieves higher accuracy than previous methods.

Keywords: gender classification, component based, multiple classifiers, occlusion image, SVM

  Retrieve PDF document (JISE_201901_06.pdf)