Abstract: Image itself has the great amount of data, it makes the transportation efficiency very low. In order to meet the demand of current network applications, it is necessary to develop a good compression algorithm which can compress the image quickly with less distortion. The paper studies two existing compression algorithms based on wavelet transformation and finds out their features and their shortages. It classifies the different wavelet trees in different frequencies after the wavelet transformation and uses a certain algorithm to compress them so that the total compression effect can be improved and the complexity can be reduced as well. Experimental result proves that the improved image compression algorithm brings better effect and efficiency than those two algorithms. Classifying and compressing respectively is effective. It improves the compression effect and reduces the complexity of the algorithm.

Key words: wavelet transformation, wavelet tree, peak signal to noise ratio

摘要： 针对图像的数据量的相对庞大、传输速度慢的问题，需要一种很好的压缩算法，既能以较少的失真率对图像进行压缩，又能使压缩的过程相对迅速，以满足当今网络应用的需求。通过研究两种已有的小波变换图像压缩算法的算法思想及算法流程，找出它们的特性及存在的不足，并通过对小波变换后的图像的不同频域子带的小波树进行分类，采用适合的压缩算法对各部分进行压缩，使图像的整体压缩效果得到提高，同时也降低了压缩过程的复杂度。实验结果表明，改进的小波变换图像压缩算法在压缩效果和压缩效率上都优于两种已有的小波变换图像压缩算法。这个分类压缩的方法能够有效地提高图像的压缩效果，也降低了算法的复杂度。

关键词: 小波变换, 小波树, 峰值信噪比