SVC Method for Textile Weave Recognizing

Support vector classification (SVC) method is proposed to dive the wrap and weft point, for recognizing the textile weave. As the densities of wrap and weft are known, this image of the textile can be divided in to rectangle of wrap and weft point. The wrap points’ histograms of R, G and B channel are different from the weft points’. So the SVC can trains the swatch by their eigenvectors which are the histograms of R, G, B channel in the rectangle selected by man, then predicts the square to display the weave of the textile. The experiences show that this method is good at to recognize the weaves from the image of the textiles which made by twist yarn or have much soft nap.