|Author:||Tsang, Ka Man|
|Title:||Study of fabric texture effects on visual and imaging color difference evaluation|
|Advisors:||Xin, John (ITC)|
|Subject:||Hong Kong Polytechnic University -- Dissertations|
Color in the textile industries.
|Department:||Institute of Textiles and Clothing|
|Pages:||xxv, 269 pages : color illustrations|
|Abstract:||Accurate color measurement plays a paramount role in establishing a reliable color quality control system in textile and garment industry. It is well-known that textile samples are rarely flat and always composed of a variety of surface textures, regarded as one of the most important parameters which influence the color appearance of the textile fabric. Such texture effect has significant impact on color quality control. Color measurement for texture textile fabrics today is widely achieved using instruments such as spectrophotometers, which average the reflection from a texture sample and may not be corrected to the visual perception. As spectrophotometer does not have spatial resolution, it fails to measure textiles and apparel with different surface structures. To date, there is not any systematical exploration of the surface texture and its influence on color appearance of the texture sample that leads to satisfactory prediction of the color difference. Recently, multispectral imaging system equipped with more than three channels is applied to estimate the spectral and spatial properties of a color fabric sample. It can accurately measure the colorimetric values, i.e., CIELAB of the fabric sample and compute the color difference for color matching applications. In this thesis, the fabric texture effects on visual and imaging color difference evaluation is investigated. The overall objective of this work is to correlate the color difference based on imaging color measurement to the visual color difference of physical texture samples. The work consists of three main parts summarized as follows. First of all, the parametric effect of fabric surface texture on color difference evaluation has been investigated via spectrophotometer and multispectral imaging system (ICM). The results show that blue texture fabrics cause the highest color difference while grey and yellow texture fabrics cause the lowest color difference. Furthermore, the total color difference for each texture pair comes from three individual color components, including lightness, chroma and hue differences in the CIELAB color space. In addition, histogram analysis was used to quantify the texture images, indicating that imaging color difference increased when the texture increased in complexity. The results from different measurement methods are greatly dependent upon the structures of the samples. Multispectral imaging system shows a higher parametric effect of texture than that of the spectrophotometer with diffuse illumination.|
In the second part, three types of color instruments such as spectroradiometer, digital camera and multispectral imaging system (ICM) were used for the instrumental and imaging color difference evaluation of woven texture samples. The color and textural on color measurement under normal and critical viewing conditions were studied by a series of psychophysical experiments using custom gray scale method. The finding also indicated that 68% of the texture pairs resulted in smaller color differences when assessed at critical viewing distance of 3.6m as compared with those viewed at 0.5m. The relationship among visual, instrumental and imaging color difference results are examined. Lastly, visual experiments were conducted to investigate the texture effect on the visual color difference evaluation at the six color centers. The imaging color differences were calculated using three models, namely, average, pixel-wise and spatial-filtering, in comparison with various color difference formulae, i.e. CIELAB, CMC, CIE94, CEIDE2000 and image difference formula IPT. The results point out that CIEDE2000 (1:1:1) color difference formula performs the best among all these formulae. The spatial-filtering model was optimized by means of optimizing parametric factors (kL, kC, kH) in the CIEDE2000 formula, Gaussian and smoothing functions, respectively. Texture features extracted from the half-width of histogram and grey level co-occurrence matrix (GLCM) were derived and used to quantitatively describe the parametric effect on the visual color difference evaluation. The combination of several texture features from GLCM was found to be significant, which led to the S-TCDM model to measure the visual color difference of two texture colors with accuracy of 11.3 PF/3 units, much better than the mean PF/3 values of 16.1 and 23.3 obtained from human observer repeatability and observer accuracy, respectively.
|Rights:||All rights reserved|
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item: