Abstract
We hoped to extract support information to enable the se lection o f the well-suited g arments to fit three-dimensional body (hereafter 3D-body) shape images o f young women. The 3 D-body simulations for evaluating the shape images were created by means o f non-tactile 3D-body measurement. Two hundred words were selected to describe various body sh ape images. Six k ey words indicating full-length body images (As : A1. Underweight, A2. Feminine, A3. Ideal, A4. Standard, A5. Masculine, andA6. Overweight figures) and 19 key words for p artial body image (Bs : B1. Leg thickness to B19. Body shape) were extracted for the classification. The 3 D-body shape images of 82 young females were evaluated o n a scale o f 1 to 5 u sing these k ey words. Six p rincipal components o f 3 D-body shape images (As and Bs) were extracted, and 7 3 D-body shape image groups (5 full-length and 2 p artial) were classified using those 5 principal component scores.
