DocumentCode
2551476
Title
Research on the embedded system of facial expression recognition based on HMM
Author
Sun, Yanbin ; An, Yi
Author_Institution
Sch. of Inf., Hebei Polytech. Univ., Tangshan, China
fYear
2010
fDate
16-18 April 2010
Firstpage
727
Lastpage
731
Abstract
With the rapid development of artificial intelligence, how to make robots have feelings has become a research highlight today. A portable device for facial expression recognition has been designed, based on the hardware platform of Intel embedded processor PXA270, and the software platform of embedded operation system Linux. The functions of video capture, image processing and face tracking have been achieved, through the application of the USB, Video4Linux programming technology, image processing technology, and artificial psychology theory. At the same time, an emotion model more consistent with the characteristics of human emotions is proposed under the theoretical framework of HMM according to the characteristics of human emotions. A new real-time facial expression recognition method is proposed based on the face detection and facial feature detection technology. Experimental results show that, the effect of face recognition is satisfactory, and the method is effective.
Keywords
Linux; artificial intelligence; embedded systems; face recognition; hidden Markov models; microprocessor chips; HMM; USB; Video4Linux programming technology; artificial intelligence; artificial psychology theory; embedded processor PXA270; emotion model; face tracking; facial expression recognition; facial feature detection technology; hidden Markov model; image processing technology; video capture; Artificial intelligence; Embedded software; Embedded system; Face detection; Face recognition; Hardware; Hidden Markov models; Humans; Image processing; Intelligent robots; Embedded System; Facial Expression Recognition; HMM; PXA270;
fLanguage
English
Publisher
ieee
Conference_Titel
Information Management and Engineering (ICIME), 2010 The 2nd IEEE International Conference on
Conference_Location
Chengdu
Print_ISBN
978-1-4244-5263-7
Electronic_ISBN
978-1-4244-5265-1
Type
conf
DOI
10.1109/ICIME.2010.5477962
Filename
5477962
Link To Document