LAMA: Stable Dual-Domain Deep Reconstruction For Sparse-View CT

TL;DR

LAMA is a novel deep reconstruction algorithm for sparse-view CT that combines data-driven and classical optimization techniques, achieving superior accuracy, stability, and efficiency in image reconstruction.

Contribution

The paper introduces LAMA, a dual-domain deep reconstruction method with learnable regularizers, integrating variational models and neural networks for improved CT image reconstruction.

Findings

LAMA outperforms existing methods on benchmark CT datasets.

LAMA reduces network complexity and memory usage.

LAMA enhances reconstruction stability and interpretability.

Abstract

Inverse problems arise in many applications, especially tomographic imaging. We develop a Learned Alternating Minimization Algorithm (LAMA) to solve such problems via two-block optimization by synergizing data-driven and classical techniques with proven convergence. LAMA is naturally induced by a variational model with learnable regularizers in both data and image domains, parameterized as composite functions of neural networks trained with domain-specific data. We allow these regularizers to be nonconvex and nonsmooth to extract features from data effectively. We minimize the overall objective function using Nesterov's smoothing technique and residual learning architecture. It is demonstrated that LAMA reduces network complexity, improves memory efficiency, and enhances reconstruction accuracy, stability, and interpretability. Extensive experiments show that LAMA significantly outperforms state-of-the-art methods on popular benchmark datasets for Computed Tomography.