The Study Of Color And Structure In The Applications Of Image Editing And Stylization

Color and structure are two important image features. They not only determine how we understand images, but also influence the artistic style of images. Color consists of the pixel values in image locations, while structure represents the variation in pixel values. Color and structure in images are closely related to each other, which is a key to image analysis and editing. In this paper, we analyze the color and structure of images,and propose several useful techniques in image editing and stylization.1. We propose a novel natural complex image colorization algorithm. Based on the analysis of the natural color formation mechanism, we find out that the colors in natural images regions lie on low rank linear space. Thus, we introduce a two-step colorization scheme that first label image regions, then colorize each region using linear interpolation. In color labeling step, users scribble color strokes to indicate different regions, the algorithm then segment the image; then in each region, users provide several color points at different gray-levels, the colors of the entire region are obtained through linear interpolation. The two-step scheme offers more intuitive interaction and better control of results for users. We demonstrate the effectiveness of our colorization algorithm using vivid colorization results in our paper.2. We introduce an interactive tool for local color transfer. The new technique is based on the observation that color transfer operations are local in nature while at the same time should adhere global consistency. We introduce a brush by which the user specifies the source and destination image regions for color transfer. Color statistics in the source region are transferred to the destination region. A global optimization is then applied to eliminate visual discontinuities that may result by the local operations. We demonstrate that our tool is easy to use yet effective in quickly generating diverse artistic effects.3. We propose a novel image stylization scheme in which image color and structure are separately stylized. This stylization scheme is applied to color sketch generation, facial pixel drawing generation and pixel architecture generation. By optimizing the image structure based on lighting distribution, our color sketch generation algorithm vividly simulate the casual hand drawing effect of sketching; In our facial pixel drawing generation algorithm, the face structure is optimized based on the facial feature importance, so that the facial feature can be best displayed within limited pixels; In our pixel architecture generation algorithm, the architecture structure is regularized to generate high quality rasterization for pixel level display. In color stylization, we introduce a color palette concept, to learn and apply artist color style to new images. This concept is successfully utilized in our image stylization techniques. A series of experimental results demonstrate that our image stylization scheme is effective in generating various art styles. The scheme is of certain value for further study.This paper introduces several effective image editing and stylization algorithms based on the study and analysis of image colors and structures. The success of these algorithms can be a useful reference for the design of future image editing algorithms.