Fractal image compression for efficient texture mapping

Abstract

Texture mapping has traditionally been used to add visual realism to computer graphics images. In the paper we propose an alternative technique for image texture handling. We use fractal image compression scheme to compress and decompress large and complex textures. We identify special properties of the method to apply in the very process of texture mapping. The main idea is to exploit high degree of local self-similarity in natural textures. The method allows for very high compression of complex textures, provides real-time decompression and perfectly suits Level of Detail. Properties of the method are compared to those of classical DCT and wavelet compression schemes.