SQUIRREL INSPIRED IMPROVED SEARCH METHOD FOR FRACTAL IMAGECOMPRESSION ON STANDARD AND MAGNETIC RESONANCE IMAGES

Abstract

The complexity in search of suitable range domain and considerable loss in
compression parameters like PSNR and MSE are the major constrictions of the baseline
fractal image compression. Hence the existing research is focused on finding optimal
solutions to pace up the search rate with marginal loss of image parameters oncompression.
Major existing fast search fractal algorithms attain the fractal search speed lowering the
image quality.Addressingthis,aSquirrelinspiredfastsearch(SIFS)method is proposed for
fractal image compression (FIC). Existing methods mainly depend
onwaveletclassification,theproposedSIFSusesmethodicalvectorofrangeblocksbased on
thesimilarity and optimizing the search based on dynamic behavior of flying squirrels
and their efficient way of gliding by the coordinate distance. The proposed SIFS method
uses foraging behavior of flying squirrels to find the best range block search showing
scalable improvements in search complexity to Particle Swarm Optimization and
Genetic Algorithm based methods. The noteworthy reduction in MSE (Mean Square Error)
calculations is observed as only six of the eight dihedral transformations are
enoughtocompare the range blocksimilarityintheproposedSIFS. Proposed method is
experimented on different kinds of images including medical MRI image and results
found are encouraging.