A Spiral Architecture Based Variable Range Fractal Image Compression Method

Abstract

Fractal Technique in Image compression is a popular method applied for locating self similarity in image through range blocks and domain blocks of rectangular shape partitions. Spiral architecture (S.A) variable range image compression fractal technique is presented here, with hexagonal shape Range blocks to find similarity for best domain block match. The presented hybrid method when applied to grey test images of different categories, demonstrates improvement in image parameters like Compression Time, Peak to signal noise ratio and Compression Time. Presented method splits the image into non over lapping range block of 7 size each and using spiral count operation searches suitable larger Domain block which is a multiple of 7.Larger domain range once located is transformed and shrink resulting in better compression and image parameters. Experimentation is performed on test standard size 512X512 images of Lena, MRI, Remote Satellite, Baboon, Bird, Rose and Pepper. Our method implemented is expected to increase performance of fractal compression parameters by maximum of 10 to 12 %.