A Novel Bit-Weight Based Transformation Selection Algorithm for IFS based Fractal Image Compression

Recently, it has been discussed in [1][2][3] the possible optimization strategies for enhancing the speed and performance of IFS algorithm for fractal image compression based on a duplicate block removal strategy. Invariably, all the existing fractal image compression algorithms suffer in performance mainly because of the lack of an efficient image classification/sorting method. A new method is proposed for further increase in the performance of IFS by implementing a seed block image sorting strategy for efficient search while selecting suitable fractal seed blocks for transformations during the IFS coding. In this method, the bit weight of an image block is derived for the individual bits of the pixel/gray values. Based on the derived weights, the image blocks are addressed and classified to reduce the search area. After selecting one image block based on their 'Bit Weight' the new block matching operation based on a mathematical model proposed. The experiments with different kinds of images have been done using this algorithm. Further Normal IFS and Bit weight based IFS image classification/sorting algorithms have been compared through time and PSNR.