Objective Evaluation of Wrinkle Recovery

摘要

A device for wrinkle recovery measurements involves the installation of a laser sensor. The shapes of fabric wrinkles are described by magnitude and sharpness using power spectra obtained from two-dimensional fast Fourier transformation and the distribution of fabric surface angles obtained from spline interpolation and vector operation. For the objective evaluation of wrinkle recovery, artificial neural networks are used rather than the AATCC subjective evaluation method. Five kinks of input patterns are categorized form the characterizations of wrinkle shape, and these are supplied to the neural networks for evaluation. Objective resluts form the nural networks are compared with subjective ones form human experts. The relationship is good with a correlation coefficient of 0.95, but wrinkle rank depends somewhat on the color or pattern of the fabric.