We propose a novel Learned Alternating Minimization Algorithm (LAMA) for dual-domain sparse-view CT image reconstruction. LAMA is naturally induced by a variational model for CT reconstruction with learnable nonsmooth nonconvex regularizers, which are parameterized as composite functions of deep networks in both image and sinogram domains. To minimize the objective of the model, we incorporate the smoothing technique and residual learning architecture into the design of LAMA. We show that LAMA substantially reduces network complexity, improves memory efficiency and reconstruction accuracy, and is provably convergent for reliable reconstructions. Extensive numerical experiments demonstrate that LAMA outperforms existing methods by a wide margin on multiple benchmark CT datasets.

翻译：我们提出了一种新型学习型交替最小化算法（LAMA），用于双域稀疏视角CT图像重建。LAMA由包含可学习非光滑非凸正则化项的CT重建变分模型自然导出，这些正则化项在图像域和正弦图域中被参数化为深度网络的复合函数。为最小化该模型的目标函数，我们将平滑技术与残差学习架构融入LAMA的设计。研究表明，LAMA能显著降低网络复杂度、提升内存效率与重建精度，并具有可证明的收敛性以确保重建可靠性。大量数值实验表明，在多个基准CT数据集上，LAMA的性能大幅优于现有方法。