Extraction of Ladies’ Wear Selection Support Information Using a 3D-Body and Garment Simulations for Adult Women's Garments (Part 4) Classification and Prediction of Design Image in 3D-Garment Designs

Abstract

In this Part 4 of our papers, the design images of sixty 3D-garment simulations (20 designs × 3 3D-body models representing women in their 20's, 40's, and 70's) were categorized into four image clusters (Al. Lady & Feminine and A5. Elegant Images (27 garments), A4. Standard & Retro Images (15 garments), A3. Casual Image (9 garments), and A2. Sharp & Modern Images (9 garments)) based on principal component scores of two factors in Part 3 of the previous paper using Cluster Analysis. The features of each design image were clarified using various of design elements (form, color, etc.), impression, clothing life, and other items. Furthermore, four design images (equations Y3: Al and A5 -A3 images and Y4: A4 -A2 images) were predicted with sufficient accuracy using design elements of garments by means of Multiple Regression Analysis. We were able to obtain useful information prior to making and selecting designs for garments using 3D-garment simulations. © 2009, The Japan Research Association for Textile End-Uses. All rights reserved.