Product shape and emotional design: an application to perfume bottles
Résumé
Shape features play a major role in the perception of designed objects. In this paper, we study the relationships between linguistic and numerical shape descriptions, focusing on real images of perfume bottles. The subjective linguistic evaluation of the bottles is obtained through expert annotations using emotional words as labels; the objective numerical description relies on automatically extracted attributes such as elongation or circularity. Statistics and machine learning tools are exploited in order to learn a matching between shape descriptors and subjective labels.