Regression-based custom skirt patternmaking using 3D body scanning data.

This study proposes a regression-based patternmaking framework to improve garment fit by accommodating body shape variation within standardised size categories. A total of 200 body measurements were extracted from 677 female 3D body scans to capture morphological diversity. The findings reveal substantial variability in body proportions within the same size category, highlighting the limitations of conventional sizing systems that rely primarily on bust, waist, and hip measurements. Linear regression models were developed to establish predictive relationships between body measurements and the coordinate-level adjustments required for a standard skirt pattern block. Results show that regression-based methods enable tailored, measurement-specific modifications to pattern coordinates for individual body shapes. The proposed modifications were evaluated through their statistical correspondence with body measurements rather than through physical or virtual garment fit testing. Accordingly, physical and virtual validation are identified as necessary future steps to evaluate improvements in garment fit.