随着多媒体技术飞速发展，各种成像设备的分辨率不断提高，一幅图像所包含的数据量越来越大。目前网络带宽受限，给图像的存储、网络的传输带来了许多困难。因此，在保证图像质量的前提下，要提高存储效率和传输速度，就必须对图像进行压缩。小波分析在时域和频域都具有良好的局部化特性，能将图像信号分解为与人类视觉特性相匹配的不同分辨率、不同方向特性的子带，并使能量集中到某些子带，为图像压缩提供了很有效的手段。 论文介绍了小波变换的相关基础理论，详细分析了常见的小波变换图像压缩算法：EZW算法和SPIHT算法，仿真比较了EZW算法和SPIHT算法的性能，结果表明SPIHT算法的性能优于EZW算法，因而选取SPIHT算法进一步研究。论文对改进SPIHT算法的改进措施、步骤做了详细介绍，仿真得出改进SPIHT算法性能总体上优于经典SPIHT算法；将9/7小波变换和改进的SPIHT算法相结合，提出了一种基于改进SPIHT算法的图像压缩方案，通过计算机仿真实验验证了此方案的良好性能。 最后，将改进的图像压缩方案应用于医学图像，仿真结果表明在速率高于0.4/bps时应用改进方案的图像PSNR值高于经典SPIHT算法且高于医学图像压缩要求34dB，效果良好，符合医学诊断要求，表明此方案可行，有医学实用价值。

With the rapid development of multimedia technology, the resolution of many imaging equipment constantly increasing, the data size of a picture is more. Because of restricted network bandwidth, the storage of image and network transmission was brought a lot of difficulties. Therefore, at the premise of image quality, to enhance storage efficiency and transmission speed, we must compress images. Wavelet analysis have good localization properties in time domain and frequency domain, the image signal can be decomposed into different resolution to match with the human visual system, different directional characteristics’ sub-band, and the energy is concentrated to certain sub-band, provides a very effective means for the image compression. Paper introduces the related basic theory on wavelet, analysis familiar image compression algorithm using wavelet transformation in detail: EZW algorithm and SPIHT algorithm, the result indicates that the effects of SPIHT algorithm is better than EZW algorithm, which decides that SPIHT algorithm is choose to research. The paper introduces the measures and steps of improved SPIHT algorithm in detail, the performance of improved SPIHT algorithm is better classic SPIHT algorithm on the whole; combined 9/7 wavelet transformation and improved SPIHT algorithm to create image compression program based on improved SPIHT algorithm, proved good performance of the program using computer simulations. The image compression program based on improved SPIHT algorithm is used in medical image, shows good effect, which directs that image compression program based on improved SPIHT algorithm has very good practical value. Simulation results show that when the rate is higher than 0.4/bps the image’s PSNR value applying the improvement program is higher than the classic SPIHT algorithm and higher than the medical image compression requirements of 34dB, the effect is good, conform the requirements of medical diagnosis, indicating that the improvement program is feasible, the improvement program has medical practical value.