サブバンド分割と組み合わせたフラクタル画像符号化

Abstract

In order to improve the performance of fractal image coding, this paper proposes a new coding scheme which is combined with resolution decomposition by sub-band. In the proposed scheme, an input image is first decomposed to low and high resolution sub-band images. The fractal block conding with adaptive range block size is performed only for the lowest resolution sub-band image. On the other hand direct quantization and entropy coding are carried out for the other high resolution sub-band image. The residual difference between reconstructed and original lowest resolution sub-band image is also quantized and entropy coded in order to enhance the coding performance. Computer simulation experiment were carried out employing 4 sub-bands and 7 sub-bands decompositions. The results show that the proposed coding scheme gives higher SNR (Signal-to-Noise Ratio) values and better reconstructed image qualities compared to the conventional fractal block coding scheme