DocumentCode :
2585372
Title :
An NARX-based approach for human emotion identification
Author :
Alazrai, Rami ; Lee, C. S George
Author_Institution :
Sch. of Electr. & Comput. Eng., Purdue Univ., West Lafayette, IN, USA
fYear :
2012
fDate :
7-12 Oct. 2012
Firstpage :
4571
Lastpage :
4576
Abstract :
This paper presents a Nonlinear AutoRegressive with eXogenous input (NARX)-based approach for human-emotion recognition from an input video. The dynamics of facial expressions are first captured by performing a temporal-spatial analysis by extracting local and spatial features using a pyramid of histograms of oriented gradients (PHOG) descriptor. Then the temporal phases of facial expressions are identified using our proposed Mutual-Information-based Delay Identification Algorithm. Finally, the emotion recognition problem is formulated into a parametric regression context using a recurrent NARX network. This approach enhances the cognitive skills of humanoid robots by adding the ability to recognize and understand affective emotional states of a human. Computer simulations were conducted to illustrate the performance of the proposed NARX-based approach with recurrent neural network realization. The proposed recurrent NARX network performed better than other existing human-emotion recognition systems and achieved a 91.5% recognition rate when tested using the Cohn-Kanade database.
Keywords :
autoregressive processes; emotion recognition; feature extraction; humanoid robots; intelligent robots; recurrent neural nets; regression analysis; spatiotemporal phenomena; video signal processing; Cohn-Kanade database; PHOG descriptor; cognitive skills enhancement; facial expression dynamics; human emotion recognition rate; human emotional states; humanoid robots; input video; local feature extraction; mutual-information-based delay identification algorithm; nonlinear autoregressive with exogenous input-based approach; parametric regression context; pyramid-of-histograms-of-oriented gradients descriptor; recurrent NARX network; recurrent neural network realization; spatial feature extraction; temporal phase identification; temporal-spatial analysis; Delay; Emotion recognition; Feature extraction; Humans; Recurrent neural networks; Testing; Training; Facial Expression Analysis; Human Emotion Identification; Recurrent NARX Neural Networks;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Intelligent Robots and Systems (IROS), 2012 IEEE/RSJ International Conference on
Conference_Location :
Vilamoura
ISSN :
2153-0858
Print_ISBN :
978-1-4673-1737-5
Type :
conf
DOI :
10.1109/IROS.2012.6385544
Filename :
6385544
Link To Document :
بازگشت