期刊
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
卷 45, 期 4, 页码 4915-4931出版社
IEEE COMPUTER SOC
DOI: 10.1109/TPAMI.2020.3012955
关键词
Convergence; Artificial neural networks; Imaging; Acceleration; Image reconstruction; Optimization; Extrapolation; Iterative neural network; deep learning; model-based image reconstruction; inverse problems; block proximal extrapolated gradient method; block coordinate descent method; light-field photography; X-ray computational tomography
Iterative neural networks (INN) are gaining attention for solving inverse problems in imaging, combining regression NNs and an iterative model-based image reconstruction (MBIR) algorithm. This paper proposes the first fast and convergent INN architecture, Momentum-Net, by generalizing a block-wise MBIR algorithm using momentum and majorizers with regression NNs. Momentum-Net guarantees convergence to a fixed-point for general differentiable (non)convex MBIR functions and convex feasible sets, under two asymptomatic conditions.
Iterative neural networks (INN) are rapidly gaining attention for solving inverse problems in imaging, image processing, and computer vision. INNs combine regression NNs and an iterative model-based image reconstruction (MBIR) algorithm, often leading to both good generalization capability and outperforming reconstruction quality over existing MBIR optimization models. This paper proposes the first fast and convergent INN architecture, Momentum-Net, by generalizing a block-wise MBIR algorithm that uses momentum and majorizers with regression NNs. For fast MBIR, Momentum-Net uses momentum terms in extrapolation modules, and noniterative MBIR modules at each iteration by using majorizers, where each iteration of Momentum-Net consists of three core modules: image refining, extrapolation, and MBIR. Momentum-Net guarantees convergence to a fixed-point for general differentiable (non)convex MBIR functions (or data-fit terms) and convex feasible sets, under two asymptomatic conditions. To consider data-fit variations across training and testing samples, we also propose a regularization parameter selection scheme based on the spectral spread of majorization matrices. Numerical experiments for light-field photography using a focal stack and sparse-view computational tomography demonstrate that, given identical regression NN architectures, Momentum-Net significantly improves MBIR speed and accuracy over several existing INNs; it significantly improves reconstruction quality compared to a state-of-the-art MBIR method in each application.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据