关键词: 乳房形态, 三维人体测量, 聚类分析, 聚拢程度, 形态判别

Abstract:

In order to study young female's breast shape discrimination for improving the bra sizing system for the Chinese females, a definition method for defining breast boundary was proposed to ensure the consistency of breast shape measurement indexes. Raw data was collected by scanning 140 female college students using the [TC]² scanner, and 28 breast measurement values including height, width, angle, and arc were extracted. Analysis of data was conducted by integrating coefficient of variation and correlation analysis methods, and 6 major parameters affecting breast morphology were identified as clustering indexes. K-means cluster was used to categorize the breast shapes into groups from the 3-D shape of the breast, and the breast shape was subdivided by the ratio of gathering degree and the young females' breast shape was divided into 9 categories. Based on the classification results on breast morphology, Fisher criterion function was used to verify the samples. The results show that the accuracy of overall judgment of the initial sample data based on morphological discrimination rules is as high as 97.1%, which shows that this criterion method has high accuracy, providing new ideas for the breast morphology research, and has a positive effect on the progress in the brassiere industry in China.

Key words: breast morphology, 3-D anthropometric measurement, cluster analysis, degree of aggregation, morphological discrimination