比较二维小波和二维小波包技术在不同压缩模式下的降噪处理

Two-dimensional wavelet versus wavelet packet technology in de-noising processing at different compression modes

背景:小波和小波包技术是进行时频信号分析的重要方法。医学图像数字化采集后断层多,数据信息量大,易受噪声影响。采用二维小波技术和小波包技术可以实现肝癌图像的完美压缩和降噪。目的:比较二维小波和二维小波包技术在不同压缩模式下压缩肝癌图像的优劣以及小波降噪的技巧。方法:选用同一幅动脉期肝癌图像,进行4层分解,利用二维小波和二维小波包技术分别进行Balance sparsity-norm、Removen ear0和Bal.sparsity-norm(sqrt)三种模式的压缩处理,再利用小波函数对含噪声信号的图像进行降噪处理。结果与结论:对于同一种压缩模式,二维小波包技术压缩肝癌图像优于二维小波技术,3种压缩模式中Bal.sparsity-norm(sqrt)模式和Remove near0 mode模式压缩比例更小,图像清晰度更好;小波降噪能很好地消除噪声信号。提示利用二维小波技术和小波包技术都可以实现肝癌图像的完美压缩和降噪。

BACKGROUND：Wavelet and wavelet packet technique is important for temporal-frequency analysis. After collection of digital medical images,they have many faults and large amount of data are sensitive to noise. The use of two-dimensional wavelet and wavelet packet technology can achieve the perfect compression and de-noising with liver image. OBJECTIVE：To compare the strengths and weaknesses of the compressed liver image using two-dimensional wavelet and wavelet packet compression technology in different modes,as well as wavelet de-noising techniques. METHODS：The same liver image of arterial phase was selected to divide into four layers. Using two-dimensional wavelet and two-dimensional wavelet packet technique,three modes of compression were performed to process the image,then wavelet function was used to eliminate the noise signal of image. RESULTS AND CONCLUSION：For the same compression mode,two-dimensional wavelet packet technology was better than two-dimensional wavelet image compression technology with liver cancer. In three compression models the Bal.sparsity-norm（sqrt） mode and Remove near0 mode＇s compression ratio was smaller with better clarity. Wavelet de-noising can well eliminate the noise signal. The use of two-dimensional wavelet and wavelet packet technology can achieve the perfect compression and de-noising with liver image.