This paper concerns with the relationship between the visual perception of the degree of pucker or wrinkles of garment surfaces and the geometrical parameters of surfaces. In this study, four potentially relevant parameters of the surface profile are considered, namely, the variance (σ2), the cutting frequency (Fc), the effective disparity curvature (Dce) (Defined as the average disparity curvature of the wrinkled surface over the eyeball distance of the observer) and the frequency component of the disparity curvature ( Dcf). Based on the experiments using garment seams having varying degree of pucker (i.e. the wrinkles along a seam line), it was found that, while the logarithm of each of these four parameters has a strong linear relationship with the visually perceived degree of wrinkles, following the Web-Fetchner Law, the effective disparity curvature ( Dce) and the frequency component of the disparity curvature (Dcf) appeared to have stronger relationships with the visual perception. This finding is in agreement with the suggestion by Rogers & Cagenello that human visual system may compute the disparity curvature in discriminating curved surfaces. It also suggested an objective method of measuring the degree of surface wrinkles.