Machine learning classification of design team members' body language patterns for real time emotional state detection. (July 2015)

• Record Type: Journal Article
• Authors: Behoora, Ishan; Tucker, Conrad S.
• Is Part Of: Design studies. Volume 39(2015)
• Page Start: 100
• Page End: 127
• Publication Date: 2015-07
• Subjects: computational models -- information processing -- design activity -- team work -- user behavior; Design -- Periodicals; 745.405
• DOI: 10.1016/j.destud.2015.04.003
• Languages: English
• ISSNs: 0142-694X

Abstract

Abstract : Design team interactions are one of the least understood aspects of the engineering design process. Given the integral role that designers play in the engineering design process, understanding the emotional states of individual design team members will help us quantify interpersonal interactions and how those interactions affect resulting design solutions. The methodology presented in this paper enables automated detection of individual team member's emotional states using non-wearable sensors. The methodology uses the link between body language and emotions to detect emotional states with accuracies above 98%. A case study involving human participants, enacting eight body language poses relevant to design teams, is used to illustrate the effectiveness of the methodology. This will enable researchers to further understand design team interactions. Highlights: Machine learning to detect emotions using non-invasive sensors in design teams. Effectiveness of approach illustrated with case study. High accuracy over 90% achieved for detecting many body language poses. Scalable solutions with potential for further research.