摘要
从优化网络结构出发,在基于迭代软阈值网络的压缩感知磁共振成像深度网络基础上,加入由p阈值函数组成的优化模块,进一步优化软阈值函数,以抑制噪声,减少重建误差,从而提高重建质量.上述算法结合了压缩感知磁共振重建和深度学习的优势,所有参数都是端到端学习得到的,既具有很好的理论可解释性,又具有良好的网络泛化能力.对上述算法与其它算法进行对比,仿真结果表明,所提算法提高了磁共振成像的重建精度,特别对于结构复杂的磁共振图像重建效果更好.
Abstract
Based on the deep network of compressed sensing magnetic resonance imaging based on iterative soft threshold network,an optimization module composed of P-threshold function was added to further optimize the soft threshold function to suppress noise and reduce reconstruction error,so as to improve reconstruction quality.The algo-rithm combines the advantages of compressed sensing magnetic resonance reconstruction and deep learning,and all parameters are learned end-to-end,which has good theoretical interpretability and good network generalization ability.Compared with other algorithms,the simulation results show that the proposed algorithm improves the recon-struction accuracy of magnetic resonance imaging.
NETL
NSTL
万方数据