A Consideration on Modeling for Perception of “Expression in Mechanics Sense”:

A Deep Learning Approach

Abstract

The design process is highly dependent on individual sensibilities in general. There is, however, also implicit objective factor in this process, such as physical rationality. In order to support the design, revealing the connection between the sensory- and the objective-factors is a key. We focused on the process providing the impression of the product shape and proposed a concept of “expression in mechanics sense” by means of the spatial pattern of stress level that might be implicitly perceived by humans. The term of “expression” is a metaphor of human facial expression. In this article, we considered a methodology for explaining how humans recognize the shape of the product through the deep learning approach.