Research And Application Of IBAR Method Based On Gradient Domain

The successes of photorealistic computer graphics in the early nineties motivated alternative techniques for rendering in non-photorealistic styles. Two decades later, the field of non-photorealistic rendering(NPR) has expanded into a vibrant area of research covering a plethora of expressive rendering styles for the visual communication. One of the most important branch of NPR is ‘image-based artistic rendering(IBAR)’, which deals with the two-dimensional content in photographs and video with artistic stylization techniques.Three kinds of innovative IBAR effects based on the gradient-domain optimization framework for image have been realized.. The three IBAR effects are claborate-style painting, freehand brush work, single blackboard and pen painting in light colors. The realization of the three effects used some key techniques proposed by Bhat et al: a robust weighting scheme, a new measure for gradient saliency, colorization and a filter for non-photorealistic rendering. We did a series of experiments on the three effects, and we improved some algorithms. During the experiments, we found a problem of colorization which based on optimization framework that the brightness of sparse color depended on the brightness of the source gray image and it limited the user’s artistic recreation. We carried out some research on this issue and put forward several solutions.We compared our experiment results and the methods of some main image processing softwares. We analyzed the strengths and limitations of our algorithm, and present some solutions to solve these problems.Finally, we studied the norms of assessing the IBAR images proposed by Hall et al. According to these norms, we assessed the IBAR effects realized in this thesis. The work of assessing the IBAR images can promote the assessment work both within and without the field and it also has the certain value to promote the academic unified the assessment norms for IBAR images.