Genetic Painting: A Salience Adaptive Relaxation Technique for Painterly Rendering
Author(s): John P. Collomosse, Peter M. Hall.
Technical Report: University of Bath, No. CSBU-2003-02, UK, October,
2004.
[BibTeX] 
Abstract:
The contribution of this paper is a novel non-photorealistic rendering (NPR) algorithm for rendering real images in an impasto painterly style. We observe that the vast majority of image-based NPR techniques operate on a local pixel neighbourhood basis, typically as non-linear image filters seeking to conserve high-frequency information in the rendering. Our work differs in that we make use of a novel perceptual salience measure, incorporating global analysis of the image. We introduce the concept of drawings as salience maps, and propose a novel painting algorithm which uses a genetic algorithm (GA) to iteratively converge toward an optimal painting in which salient detail is conserved and non-salient detail is attenuated. Differential rendering styles are also possible by varying stroke style according to the classification of salient artifacts encountered, for example edges or ridges. We demonstrate the results of our painterly technique on a wide range of images illustrating the benefits of rendering preferentially with regard to salience, and the improvements gained by subsequent relaxation of the painting using our GA technique.