Zooming using iterated function systems

Iterated Function Systems (I.F.S.) have been recently studied in the field of image coding. In addition to compression, I.F.S. possess some properties of fractals which can be used to sub/oversample an image. In this paper, we review how I.F.S. can be used for image zooming, directly from basic algorithms of still image compression. This study shows that results obtained in such a way do not produce better results than those obtained by using classical spatial interpolators such as the linear one. However, in this paper we also show that modified versions of the basic approach provide a better use of I.F.S.'s coding for zooming. 1. INTRODUCTION Since 1990, I.F.S. have been studied in the field of image coding. The reference algorithm, from Arnaud Jacquin, is briefly reviewed in section 3. Since then, several relevant papers proposed some improvements. Moreover, some authors indicate that the fractal code built during the coding stage is independent of the size of the original.