Entropy Based Lossless Fractal Image Compression using Irregular Rectangular Partitions

Entropy of an image can be taken as a parameter of variation among pixel values. Equal value for all pixels in an image results in zero entropy. This idea is incorporated at the time of partitioning the image. Partitions are done with zero entropy in order to make the compression lossless. Unlike traditional fractal image compression mechanism this method doesn’t require two separate partitions called range blocks and domain blocks on same image, rather it needs only single irregular rectangular partition. Each pixel compares its value with adjacent pixels and as long as their values are same, the pixels remain in same partition. The fractal approach is incorporated by reading the common pixel value in a partition and searches other partitions with same pixel value in order to store the block once. The proposed method increases encryption and decryption speed considerably over traditional methods. KeywordsEntropy, lossless compression, fractal, domain block, range block.