In order to evaluate the puckering of seams on three-dimensional garment su
rfaces objectively, the surface contours due to garment silhouette and drap
e must be taken into consideration. In this paper, a 2-D FIR band-pass digi
tal filter is proposed to extract pucker profiles from the measured profile
s by removing the 'high-frequency' components due to fabric-surface texture
and 'low-frequency' components due to three-dimensional garment-surface co
ntours. The filter was found to be very successful when applied to evaluate
seams, even when twisted and laid on double-curvature surfaces that simula
te the complex three-dimensional garment surfaces. The objective parameters
derived from the pucker profiles after 2-D filtering were found to be line
arly correlated with the subjective AATCC pucker grades.