Emotional state representation and detection method of users in library space based on body posture recognition

Abstract

Purpose In the information commons (IC) space of library, it is very important to recognize the emotional state of users for better playing the role of IC. In view of this point, this paper aims to discuss the human expression of user emotion. Design/methodology/approach An emotional state recognition method based on body posture change under video monitoring is proposed. In this method, two parameters are proposed to represent the emotional state of users. Finally, the distribution of users’ overall emotional state is recognized. Findings It is found that the change of human posture reflects the emotional state of users to a certain extent. The spatial frequency of the user’s average body position change and per capita body position change can reflect the spatial distribution of individual and body position change, respectively. Originality/value The method in this paper can effectively overcome the inaccuracy of manual identification of video monitoring images, especially in the case of a large number of users and effectively help the construction of university library IC space and provide a basis for the setting of environmental parameters.