基于小波极大模值信息的医学图像去噪方法

A technology of denoising medical image based on wavelet maximum module information

目的利用小波变换进行医学图像去噪。方法通过分析二进小波变换下小波极大模值的特点,即信号的极大模值往往会大于噪声的极大模值,而且噪声的极大模值会随着尺度增大而急剧减少,信号的极大模值却改变很小,由此构造了更有效的去噪准则,即根据不同尺度上的极大模值信息,选择不同的域值来滤除噪声。结果应用该方法进行医学图像去噪,能保持较高的峰值信噪比、图像细节和边缘特征以及图像清晰度。结论基于小波极大模值信息的去噪方法能有效地降低医学图像中的噪声。

Objective Apply wavelet transform to denoise medical image. Methods This paper analyzes the characteristics of wavelet maximum module using binary wavelet transform, and finds that the maximum module generated by signal is always higher than that by noise, the maximum module generated by noise decreases rapidly when scales increase, but maximum module generated by signal changes very little. So we form a more effective denoising principle, which is to choose different thresholds according to the maximum module information on different scales. Results If we apply this method to denoise medical image, it can get a better peak value of signal-to-noise radio, image detail, edge characters and definition. Conclusion The denoising method based on wavelet maximum module information can reduce the noise in medical image effectively.