Abstract:

In order to establishing quantitative relationship between breast surface area and bra pattern-making parameters, young non-pregnant female students were scanned by 3-D body scanner, and breast feature points and boundary line were defined and extracted by reverse engineering software, then local breast coordinate system was built based on the feature points. According to the horizontal and segittal plan of local coordinate system, breast surface was divided into 4 parts as upper-inside, upper-outside, lower-inside and lower outside, and surface area of each part was measured. Body measurements related to bra pattern making were selected, regression models between body measurements and surface area of each part were established by the use of stepwise partial least squares regression method. Upon examination, no significant difference exists between the original values and forecastiong values of each breast surface area, and all the models have high precision, and can provide efficient breast surface area predicting method for breast shape analysis and bra personalized customization.

Key words: breast, reverse engineering, body measurement, partial least squares regression, stepwise regression