Embedded Wavelet Image Compression Algorithm Based On Full Zerotree

In this paper, a novel embedded wavelet image compression algorithm, NEFZW (Novel Embedded Full Zerotree Wavelet) algorithm is proposed. The relationship of the intra scale subbands is exposed by the statistical experiments on various test images. Since “zerotree” indicates specially the inter scale relationship in EZW, the concept of full zerotree is proposed which includes both the relationship of inter and intra scale. To make NEFZW feasible for practical coding, the idea of “vector pair” is introduced into NEFZW to improve the coding efficiency; meanwhile, an additional symbol is utilized to avoid the duplication of insignificance coding among intra scales. Based on these, NEFZW algorithm is described. And the corresponding experimental results show that NEFZW can save the symbols in the significance map by more than 20% without otherwise affecting the performance.