Evaluation of Visual "Cool Feel" of Textile Design Using Wavelet Transform.

Abstract

There are subtle variations of density values on textile surface images. It is difficult to correctly analyze the subtle differences without human subjective evaluation. But the textile surfaces generally have a kind of periodicity. So it was supposed that irregular parts of frequency components on the textile surfaces are related with the visual features of the textiles. So in order to extract the irregular parts, the wavelet transform method that can be obtained local information was introduced. In this research, we proposed the method of objective evaluation of visual “cool feel” of textile designs using the wavelet transform. First, the warp and weft yarn's cycles in the textile images were investigated by applying the fast Fourier transform (FFT) to their images. Second, “mother wavelet” functions were constructed by using these cycles, and the wavelet transform with the “mother wavelet” functions was applied to the original images. Third, the wavelet-transformed images were converted into binary images and obtained information was regarded as the objective evaluation scale. Finally, the validity as one method of objective evaluation is confirmed by investigating correlation with subjective evaluation.