Construction and Psychological Evaluation of 3-D Emotion Space

FUMIO KAWAKAMI, HIROSHI YAMADA, SHIGEO MORISHIMA, HIROSHI HARASHIMA
1995 International journal of biomedical soft computing and human sciences  
AND HIROSHI HARASHIMA4 Seikei Uhiversityi, Kbwamura Cbllqge2, U)iiversity of lbronto3, Uhiversity of lbkyo4 (]leceived 9 lanuary 1995; revised and accepted 25 Mdrch 1995) Absmact: lhe goal of research is to realize namrat huntan-machine communication system by giving a focial expression to computer. Ib realize this environmeng it ls necessary for the machine to recognize humanls emotion condition appearing in the httmanls facq and synthesize the reasonable fociat image of machine. Iibr thts
more » ... ose, the machine siiould have emotion model based on parameterized faces which can express an einotion condition quanthatively. Especially, both mapping and inverse mapping frDm foce to the model have to be achieve, FZicial au,ression is parameterized with IibcialAction Cbding System(E4CS) mahich is translated to the grids'motion of foce model. An emotion condition is described compactly by the point in a 3D space generated by 5-tayered neural nemork and its evaluation resutt shows the high pet:fonmance of this modeL Kleyword: foce mode4 emotion mode4 neural nemor:k, computer graphics nonverbal communicationg E4CSI six basic expressions 1. Introduction Facial expression is including seveTal kinds of factors which can express non-verbal information. So Tecognizing and synthesizing the facial expression can suTely improve the communication environment between human and machine. Emotion is one of the most irnportant factors which facial expression can descTibe. Especialiy we focused on this emotion condition which appears on humanls face. OuT goal is to realize very natural hurnanmachine communication environment by giving a faoe to computer terrninal or communication system Of course, this system has to be able to recognize the
doi:10.24466/ijbschs.1.1_33 fatcat:2l6odkans5esjeknswuxvoqnlq