3차원 인체 형상을 이용한 길 원형 알고리즘 개발

Abstract

The aim of this study was to develop a new planarization technique of body surface to use in classifying clothing and body types by analyzing 3D virtual human body shape. Previous studies on planarization of body surface focused on forming a mesh on the surface of 3D shapes, spreading it on a 2D plane and then applying it. A number of constraint conditions and rules were required so that the shape of planarized body surface could be similar to the original shape of clothing. In addition, it was not easy for the planarized shape to be used for clothing. Therefore, this study developed the planarization technique of body surface close to the shape of basic bodice pattern not by planarizing through the spread of mesh pieces but by maintaining the body surface area of the targeted 3D shape and length of major structural lines. For this purpose, after analyzing the principle of generating clothing patterns from the human body in clothing construction, it should be used as a basic principle of planarization in this study. Primarily, fundamental considerations of human body's reference points and structural lines to reflect the characteristics of the human body in patterns and divided area were required. This study analyzed the characteristics of musculoskeletal form to apply them to the planarization patterns, and the drawing of planarization patterns was made by adding the structural lines to the basic line of basic bodice pattern in the same way as the drawing method of basic bodice pattern. The most important factor in the basic bodice pattern is to complete patterns by maintaining the length of structural lines and divided area of the human body as much as possible. Therefore, the fundamental structural lines and divided area of the shape measured by the 3D virtual program created a calculation formula converted by the coordinate point of body surface planarization patterns based on the drawing method of basic bodice pattern. Due to the complex mathematical calculation, this study additionally developed a program that calculated to convert into the coordinate points of planarization patterns when entering the length and width of the designated structural lines of 3D virtual human body shape. The algorithm and program developed in this study can be exclusively used for women with a normal BMI (18.5 to 29.99). This is because this study set the control point, dart and structural lines by extracting the maximum protrusion by musculoskeletal form based on the principle of drawing of basic bodice pattern. The maximum protrusion varies depending on the excess body-fat distribution other than the musculoskeletal characteristics among women with overweight and obesity. It is not suitable for the drawing principle in this study. The three 3D virtual shape models of women with a normal BMI were applied to the developed algorithm. As a result, the structural lines of the body surface pattern completed by using the coordinate points finally output showed an error within approximately 0.25% compared with the measured value of 3D shape, the divided area showed an error within 1.8% in the area including the curved part and within 0.018% without the curved part. Compared with the results of previous studies, length error of 1 to 2 % and area error of 1 to 3%, it is possible to confirm that the deformed length and area substantially reduced. However, when constructing the calculated body surface patterns directly to clothing, expressions of the armholes and neckline curve were inappropriate, and it was not appropriate to use the ease given by the previous clothing patterns in transforming the body surface patterns to clothing patterns. Therefore, in order to use the body surface patterns developed in this study for clothing, it is necessary to develop pattern curve equation with a more natural form, and additional studies on ease setting are required to use for clothing patterns directly. Moreover, extensive studies on additional morphological characteristics of the human body such as body-fat distribution should be added to this algorithm so that they can be used for overweight and obesity.