| Summary: | 碩士 === 國立臺灣科技大學 === 高分子系 === 94 === This thesis proposes a computerized color separation system for printed fabrics by using an approach of artificial neural network. This system can be applied accurately in color separations that use a scanner to obtain digitized color images in the RGB (Red, Green, Blue) mode. Next, a Genetic Algorithm (GA) is applied to search for smaller sub-images with the same color distribution of the original printed fabrics in order to proceed with the subsequent color separation algorithm. In respect to the color separation algorithm, this system can be operated by two supervised learning networks on the RGB image of the printed fabrics. First, the color separation is conducted by using the Back-propagation Neural Network (BPNN) on the RGB sub-images of the printed fabrics. Second, it is conducted by using the Probabilistic Neural Network (PNN) on the RGB sub-images of the printed fabrics. Finally, as for the color matching algorithm, it is proceeded with the color ticket of the PANTONE Textile Numbering System. According to the experimental results, these two kinds of color separation systems can successfully complete color separations and color matching for printed fabrics' images, and the PNN method is more suitable for color separations of printed fabrics.
|
|---|
