【摘要】 针对连续存在的零树根的现象,提出了一种改进的嵌入零树小波图像编码(embedded zerotree waveletsimage coding,EZW)映射算法,对于子孙均为次要系数的重要系数,用新的标志位代替其子孙中大量连续出现的零树根,对标志位采用哈夫曼编码,再将编码后的标志位比特流与幅值量化得到的比特流组合进行游程编码。针对小波系数成簇分布的特点,首先提取重要系数,然后采用形态学膨胀操作保留其边缘和纹理信息。实验结果表明,新算法极大地减少了零树根的数量,节省了扫描时间,在不影响重构图像质量的同时提高了压缩比和编码效率。更多还原

【Abstract】 An improved EZW for the phenomenon of successive zerotree root was proposed.For those descendants of a significant coefficients were all insignificant,the new symbols were used to replace of the successive zero tree roots of frequent appearance.The Huffman coded symbols were then Run-length coded to recode the symbol′s binary stream and the quantization of the amplitude.The significant coefficients were extracted,then the edge and texture information were reserved using morphological expanding operation for the characteristics of clusters of wavelet coefficients.Experimental results showed that this method greatly reduced the number of zero tree roots through re-classifying the significant symbols,saved the scanning time,and promoted compression ratio and coding efficiency without affecting the quality of reconstructed image.更多还原