Abstract:

The term fractal is used to designate objects that are self-similar and have details at different scales. Fractal Dimension (FD) is a measure to quantify how densely fractal occupies the space in which it lies. This characteristic has been used in texture classification, segmentation, clustering, edge detection and other problems. Local fractal dimension (LFD) is the FD value of a local area rather than the whole image. In this paper the concept of LFD is applied is applied in fractal image compression. Conventional iteration-free fractal image compression scheme requires a large amount of  encoding time for performing range-domain mapping. It is decreased by classification of image parts based on their local fractal dimension. First, a preprocessing analysis of the image identifies the LFD value of each range and domain block. Then, domain blocks are clustered into three groups. Range domain mapping is made on any one of the three clusters. This proposed method reduces the compression time by about 80% when compared with conventional iteration-free fractal image compression scheme at the same time maintaining the image quality.

Keywords: Local Fractal Dimension, Fractal Compression, Domain Pool, Affine Transformations.

BibTex:

@ARTICLE{P1150619005,

AUTHOR = {M.L.Valarmathi and K.Anbumani},

TITLE = {Non Iterative Fast Fractal Codec Using Local Fractal Dimension},

JOURNAL ={ICGST International Journal on Graphics, Vision and Image Processing},

}

(Full Paper 458KB)