Wavelet-Based Adaptive Embedded Fractal Image Coding

Abstract

An embedded fractal coding scheme for low bit rate transmission of discrete wavelet transform coefficients is proposed in this paper. Using 3-level 2D discrete wavelet transform (2D-DWT) for image data compression, fractal encoding is employed for generating the fractal parameters of the DWT coefficients of the third level subimages (the coarser scale). This should decrease the encoding fractal processing time. The subimages of the third level and that of the subsequent levels (finer scales) are reconstructed from the transmitted fractal parameters of level three by exploiting the high correlations among the DWT coefficients in different scales which is called embedded fractal coding (EFC) scheme. For further reconstructed image quality improvement, an adaptive technique using vector quantization is used to transmit some DWT coefficients. This is called the adaptive embedded fractal coding/vector quantization (AEFC/VQ) scheme. Experimental results showed that the proposed scheme can provide, with low processing time, good image quality in terms of peak signal to noise ratio (PSNR) at low bit rate