Design Support of function of brassiere cup using gaussian process regression

Abstract

A method to design the function of the brassiere cup shape as developable surfaces and its developed shape using Gaussian Process Regression is proposed. A developable surface, which is generated by sweeping a straight line along a three-dimensional curve, can be seen many products such as ships, buildings, clothes, and so on. The shape has not only its aim which can be formulated but also that which cannot be formulated such aesthetics. In this paper, we focus on a brassiere cup. A brassiere cup is composed of several patterns and the cup shape is designed by repeatedly making paper cup model and then checking its three-dimensional shape. For improvement of design efficiency of brassieres, such trial and error must be reduced. The difficulty of the design process is caused by the function of a brassiere cup. Its function, such as to enhance woman’s breast size, et.al., is difficult to formulate and unclearly correlated with its three-dimensional cup shape. In this paper, we propose a method to support the design of the three-dimensional shape of a cup and its developed shape by machine learning when the cup shape and quantitatively evaluated value of its function are given as a set of data. First, we formulate the cup shape as developable surface using differential geometry. Then, we propose the method to extract the attribute from the three-dimensional cup shape based on the differential geometry and a predictor of an output value for its attribute using Gaussian Process Regression. The validity of the method is confirmed by a numerical experiment regarding the evaluated value using its volume and size. Finally, we propose a method to design the cup shape using this predictor. We experimented whether our proposed method can output the approximate cup shape when the evaluated value of the cup is given.

Keywords

• design
• modeling
• developable surface
• machine learning
• gaussian process
• differential geometry