期刊名称：International Journal of Advanced Research In Computer Science and Software Engineering
印刷版ISSN：2277-6451
电子版ISSN：2277-128X
出版年度：2012
卷号：2
期号：5
出版社：S.S. Mishra
摘要：Facial expre ssion recogniti on h as attracted much atten tion in recent years becau se of it s im portance in realizin g hi ghly in telli gent h uman-machine in terfaces. I n thi s paper, we propose a new facial expression recognition technique th at utilizes the 2 -D DCT of full size facial images and ei genvector s based feed fo rward neu ral n etwork. The neu ral n etwork is designated to separate group of facial expressions with members "happy", " surp rise", "an ger", "sh ock", "di sgust" and "sadnes s". This work is the fi rst trial to us e feed forward neural network, eigen vectors an d 2-D DCT simultaneously with in a sin gle recogn ition task. To demonstra te the capability of the propos ed re cogn ition tech niqu e, we use two databases, inclu ding a recently con structed on e, which contain 2 -D front face images of 120 men and 80 women , respectively. Expe rimental resul ts reveal th at th e new technique outperform s, on the whole, the si mple vector match ing and K-means based vector matchi ng techniques and two recently de veloped m ethods u sing fixed size and constructive neu ral n etworks. The mean recogniti on rates of the new tech niqu e h ave been found as hi gh as 98.5% and 95.8% for these two database s, re spec tively, wh ich are the best resu lts among those obtained for the sa me databa ses
