一种彩色图像的小波域预测编码方法

彩色图像存在空间相关性和帧间相关性,如何有效的去除这两种相关性是提高彩色图像压缩比的关键。提出一种小波变换域的自适应预测无损压缩方法,分为三个步骤:首先对彩色图像各个分量进行可逆整数小波变换减小空间相关性,然后利用联想记忆神经网络对小波系数图像分别进行空间预测和帧间预测,进一步去除空间和帧间相关性,最后对预测残差进行算术编码。实验证明该算法是可行的和有效的,对测试图像的压缩效果优于JPEG LS标准。

Prediction coding based on wavelet transform for color image lossless compression

Redundant information can be observed in color image, which include intra-color correlation and inter-color correlation, so decorrelation is the key for color image compression. An adaptive prediction algorithm based on wavelet transform is proposed, which can be divided into three steps: first, integer wavelet transform is used to resolve each components individually to weaken spatial correlation; then coefficients of wavelet transform are adaptive predicted based on Associative Memory System to reduce spectral and spatial correlation; at last arithmetic coding is exploited to code the errors between predictive value and coefficients. Experiments give better results than JPEG-LS and show that the method is effective for image compression.