一种基于IWT与自适应预测的无损压缩方案

Lossless compression scheme based on IWT and adaptive prediction

提出一种基于IWT与自适应预测相结合的无损图像压缩方案IWTAP。详细阐述了该方案的理论依据,在此基础上,结合矩阵理论和低频子带的自相关模型,给出了自适应预测参数的简化计算公式。编码时,每次IWT后,用低频子带改进高频子带一次,改进时所需的参数依据低、高频子带的特征自适应确定;解码时,先进行高频子带还原,再进行逆IWT。实验结果表明,该方法能有效降低高频子带的信息熵,降低图像的无损压缩比特率,计算复杂度上升不大。

A novel lossless image compression scheme IWTAP was presented based on combining integer wavelet transform and adaptive prediction. The theoretic arguments of the scheme was clarified first, and then the less complicated formula were given for computing these parameters which were deduced according to matrix theories and self-correlation model of low pass. At encoder side, high pass coefficients were modified by low pass coefficients following each time integer wavelet transform. Parameters used in modification were produced adaptively based on the characteristics of low and high pass coefficients. At decoder side, the high pass coefficients were reverted firstly, and then the integer wavelet transform was reversed. Experiments show that the scheme can greatly decrease the entropy of high passes and the lossless image compression bit rate, while the computational complexity increases slightly.