Study of human perception on body shapes using intelligent techniques

Abstract

This paper presents a method permitting to determine the relationship between body measurements and human perception on morphology of body shapes, expressed by a set of emotional descriptors and fashion themes. It can be integrated into current garment CAD system for providing personalized design support. Two procedures of sensory evaluation have been used. The first one aims at evaluating abstract fashion themes by means of emotional descriptors. In this procedure, the evaluators, including fashion designers and consumers, will give relevancy degrees between fashion themes and emotional descriptors. A fuzzy cognitive map is introduced to represent the corresponding results. It can be considered as a conceptual model characterizing the cognition level of evaluators on abstract fashion themes. The other procedure aims at evaluating the previous emotional descriptors on human body shapes according to the corresponding body measurements. A set of decision trees have been established to model the relation between the emotional descriptors and the human body measurements. It can be considered as a empirical model obtained from learning data.