Fractal Image compression is based on existence of self-symmetry in image. It is a very advantageous technique in the field of image compression. The computational expenses of suitable domain search make it very time consuming in encoding part. A lot of advances have been made, but the high encoding time still remains the drawback of this technique. In this paper we have proposed a new suitable domain search technique based on feature extraction. Suggested scheme reduces the number of range-domain comparisons with significant amount and gives improved time performance. We have investigated the outcomes of image size variations on its performance.
Feature vector, domain block, range block, suitable domain search, binary tree, speed-up
