【摘要】 将基于上下文的算术编码方法和位平面编码方法有机结合，得到了基于小波包变换的分层预测图像压缩算法 Ｂ Ｐ Ｐ．为了充分利用小波系数间的各种相关性，在组合父子、邻居和兄弟相关性基础上得到了其它四种相关性．提出的小波包基选择方法既保持了父子相关性，又简单有效．根据三类位数据的特点，进行了相应的预测和编码．对比分析表明 Ｂ Ｐ Ｐ的压缩效果优于已有的其它小波图像压缩算法．更多还原

【Abstract】 By applying the bit plane coding and the context based arithmetic coding to wavelet domain, a new method, the bit plane predicting image compression algorithm based on wavelet packet transform (BPP), is obtained in order to exploit various kinds of correlation among wavelet coefficients thoroughly, other four kinds of correlation were defined by combining the parent children, neighborhood and siblinghood correlations. The selection method for wavelet packet bases not only remained the parent children correlation, but also was simple and effective. Three classes of bits were predicted and coded respectively according to their properties. Experimental results show that BPP performs better than other existing wavelet image compression algorithms.更多还原