Fractal Coding Using Gradient Direction Based Tag Matrix and Score Value

Reducing encoding complexity and improving the tradeoff between quality of image and compression ratio are active areas of research in the field of fractal image compression. In this paper two techniques are proposed to reduce the encoding complexity and to improve the PSNR of the reconstructed image. The first method is the prediction of affine transformation, based on the gradient direction of domain and range block using a Tag Matrix and the second involves calculation of score value based on the maximum value of intensity difference of range and domain block. An important merits of this new method is that encoding time is highly reduced and PSNR is also improved than the existing fractal coding techniques. This technique will be very applicable in situation where we have limited storage space and bandwidth but have to store and transmit enormous amount of data.