Tie-dye technique and pattern features

Abstract

Relationship between tie-dye technique and image pattern has been studied. Based on digital image processing, average value of HSV (hue, saturation, value) tri-component of valid tie-dye area, proportion of incompletely dyed area in HSV color space and the Tamura first three texture features of digital tie-dye image have been extracted. Then, the repeated tests and analysis of variance (ANOVA) of rotation speed and concentration have been done. The results show the obvious impact of concentration and rotation speed on the color and texture feature of tie-dye pattern. Furthermore, back propagation neutral network is developed and partial pattern features with low correlation is used to forecast the tie-dye concentration and rotation speed. The experiment results show the 100% rate of forecast, which proves that pattern features can effectively achieve the technique forecast.