SENSORY EVALUATION AND VISUAL IMFORMATION OF COLOR TEXTURE AND FABRIC APPEARANCE USING IMAGE ANALYSIS

点击翻译按钮获取中文摘要

Gray scale image analysis is applied for the measurement of visual features of plain knitted fabrics with different kinds of stitch densities, crease appearance of regularly or irregularly folded cotton cloth, and visual relaxation behavior of randomly wrinkled cloths including wrinkles like ripple and irregularly folded creases. Mean of the gray level for the digital image, the angular second moment, correlation, contrast and entropy which were extracted from the gray level co-occurrence matrix, and the fractal dimension are measured as visual feature parameters. These image information parameters is found to be useful for characterizing the beauty of plain knitted fabrics, the visual features of creased cloths and the visual relaxation behavior of wrinkle recovery.Color Image analysis is applied to the characterization of color contrast and colored shadow phenomenon of polka-dotted patterns. Colors in an image are expressed in the NTSC_RGB color system. RGB images are converted to L* … More C*H* images. The co-occurrence contrast (CON) is calculated for L*, C* and H* images. CON is found to be a good parameter to describe the contrast effect and colored shadow of polka-dotted patterns. Color image analysis is used to evaluate the warmth and coolness of multicolored patterns of cloth from a whole color impression of multicolored patterns by applying the logic of juxtaposition color mixture. It is concluded that warm, neutral and cool color images of multicolored patterns of cloth can be evaluated by hue angles. Color feature analysis by using RGB color space is applied to the characterization of color textures of many fabrics. The relationship between their sensory evaluation values and the visual feature parameters is examined by the factor analysis. In color image analysis of color fading for color patterned fabrics, L*, a* and b* are computed for each pixel in the image using the CIELAB formula. Mean values of L*, a* and b* for all pixels in the image are obtained as AVE-L*, AVE-a* and AVE-b*, respectively. There is a good agreement between L*, a* and b* by spectrometer measurements and AVE-L*, AVE-a* and AVE-b* by image analysis for solid color. Less