基于压缩感知的医学图像重建方法

Compressed Sensing-based Methods for Medical Image Reconstruction

压缩感知理论广泛应用于信号和图像处理当中,该文分别利用三种基于压缩感知的重构算法,即凸优化算法中的基追踪算法、贪婪算法中的正交匹配追踪和分段正交匹配追踪算法,在采用小波变换实现稀疏表达的基础上,比较其二维图像重构效果。实验结果表明,借助压缩感知理论,能借助少量的稀疏系数,来精确重构出原始图像,图像质量与原始图像差异不大。因此,基于压缩感知的医学图像重建方法,在医学图像处理领域,具有十分重要的理论意义和临床应用价值。

Compressed sensing theory has been widely used in signal and image processing.In this paper,three reconstruction algorithms based on compressed sensing were used,namely,the base tracking algorithm of theconvex optimization algorithm,the orthogonal matching tracking algorithm of the greedy algorithm and the piecewise orthogonal matching tracking algorithm.Based on the wavelet transform to achieve sparse representation,the two-dimensional image reconstruction effects were compared.The experimental results showed that the original image could be accurately reconstructed based on a small number of sparse coefficients with the help of compressed sensing theory.The quality of the reconstructed images showed no obvious difference with that of the original images.Therefore,it suggests that the medical image reconstruction method based on compressed sensing has not only great theoretical significance but also great potential for medical image processing in clinical applications.