返回Google圖書搜尋

Generation of Emotional Feature Space for Facial Expression Recognition Using Self-Mapping

出版	INTECH Open Access Publisher, 2010
ISBN	95330707499789533070742
URL	http://books.google.com.hk/books?id=i1ytoAEACAAJ&hl=&source=gbs_api

註釋The study described in this chapter examined the method of generating a subject-specific emotion feature space for estimating the grade of emotion. The essential results obtained in this chapter are the following. 1) Hierarchical use of SOM with a narrow mapping space enables extraction of subjectspecific expression categories. 2) The grade of emotions with "pleasantness" and "arousal" as indices can be matched to the grade of change of facial expression patterns on an EMap that is generated using the proposed method. 3) An EMap enables quantification of the grade of emotion to the grade of change of facial expression patterns, and to conduct emotion estimation to mixed facial expressions. Objective evaluation on the relationship between the temporal change of facial expression patterns and state transition on an EMap accompanied by the context of a scene will be performed in a future study using natural facial expressions during conversation.