Abstract

We integrate stylized rendering with an efficient multi-resolution image representation. This enables the user to control how compression affects the aesthetic appearance of an image. We use a point-based rendering approach for progressive image transmission and compression. This represents an image by a sequence of colour values. To best approximate the image at progressive levels of detail, we developed a novel, adaptive farthest point sampling algorithm that balances global coverage with local precision. We store no spatial information other than the aspect ratio: the spatial position of each colour value is inferred from the preceding members of the sampling sequence. We keep track of the spatial influence of each sample on the rendition, and use a progressively-generated discrete Voronoi diagram as the common foundation for our sampling and rendering framework. This framework allows us to extend traditional photorealistic methods of image reconstruction by scattered data interpolation to encompass non- photorealistic rendering. It supports a wide variety of artistic rendering styles based on geometric subdivision or parametric procedural textures. We compare our results with standard JPEG compression and conclude with a brief overview of the implications of using non-photorealistic representations for highly compressed imagery. This work was undertaken largely by Dr Mark Grundland (Dr Dodgson’s PhD student). Part of the work formed a Part II project undertaken by Mr Chris Gibbs.