Abstract
This study is an attempt to use 3D anthropometrical database for the manufacture reference of dress form for female attire. The database consists of 500 samples of female whose age is between 20 to 40 years old. Each sample consists of about 200,000 body surface points that are measured by a 3D body scanner. The proposed method consists of sizing strategy and 3D form analysis method. The sizing strategy is to classify the database into some groups based on waist circumference as the primary classification dimension and stature as the secondary. We use two steps "form difference analysis" method to extract an ideal representative human model out of each size group. The first step is to extract six representative features. They are front chest, back chest, front waist, back waist, front buttock and back buttock. The second step is to compare the six features between the entire sample and the six representative features. The sample person who has the smallest "group difference" will be the representative human. The form difference analysis method is basically based on the idea of the area difference of the segments between two individual samples. This “individual difference” is the summation of the area difference of the correspondent cross-section of these two samples. By comparing one individual sample with the rest of samples in the group and summing up all the individual difference, we derived “group difference”. The sample that bears the smallest group difference within the size group is chosen as the representative sample for that group.