On instabilities of deep learning in image reconstruction and the potential costs of AI.

Deep learning, due to its unprecedented success in tasks such as image classification, has emerged as a new tool in image reconstruction with potential to change the field. In this paper, we demonstrate a crucial phenomenon: Deep learning typically yields unstable methods for image reconstruction. The instabilities usually occur in several forms: 1) Certain tiny, almost undetectable perturbations, both in the image and sampling domain, may result in severe artefacts in the reconstruction; 2) a small structural change, for example, a tumor, may not be captured in the reconstructed image; and 3) (a counterintuitive type of instability) more samples may yield poorer performance. Our stability test with algorithms and easy-to-use software detects the instability phenomena. The test is aimed at researchers, to test their networks for instabilities, and for government agencies, such as the Food and Drug Administration (FDA), to secure safe use of deep learning methods.

Convolutional Analysis Operator Learning: Dependence on Training Data.

Convolutional analysis operator learning (CAOL) enables the unsupervised training of (hierarchical) convolutional sparsifying operators or autoencoders from large datasets. One can use many training images for CAOL, but a precise understanding of the impact of doing so has remained an open question. This paper presents a series of results that lend insight into the impact of dataset size on the filter update in CAOL. The first result is a general deterministic bound on errors in the estimated filters, and is followed by a bound on the expected errors as the number of training samples increases. The second result provides a high probability analogue. The bounds depend on properties of the training data, and we investigate their empirical values with real data. Taken together, these results provide evidence for the potential benefit of using more training data in CAOL.

Efficient Compressed Sensing SENSE pMRI Reconstruction With Joint Sparsity Promotion.

The theory and techniques of compressed sensing (CS) have shown their potential as a breakthrough in accelerating k-space data acquisition for parallel magnetic resonance imaging (pMRI). However, the performance of CS reconstruction models in pMRI has not been fully maximized, and CS recovery guarantees for pMRI are largely absent. To improve reconstruction accuracy from parsimonious amounts of k-space data while maintaining flexibility, a new CS SENSitivity Encoding (SENSE) pMRI reconstruction framework promoting joint sparsity (JS) across channels (JS CS SENSE) is proposed in this paper. The recovery guarantee derived for the proposed JS CS SENSE model is demonstrated to be better than that of the conventional CS SENSE model and similar to that of the coil-by-coil CS model. The flexibility of the new model is better than the coil-by-coil CS model and the same as that of CS SENSE. For fast image reconstruction and fair comparisons, all the introduced CS-based constrained optimization problems are solved with split Bregman, variable splitting, and combined-variable splitting techniques. For the JS CS SENSE model in particular, these techniques lead to an efficient algorithm. Numerical experiments show that the reconstruction accuracy is significantly improved by JS CS SENSE compared with the conventional CS SENSE. In addition, an accurate residual-JS regularized sensitivity estimation model is also proposed and extended to calibration-less (CaL) JS CS SENSE. Numerical results show that CaL JS CS SENSE outperforms other state-of-the-art CS-based calibration-less methods in particular for reconstructing non-piecewise constant images.

Efficient compressed sensing SENSE parallel MRI reconstruction with joint sparsity promotion and mutual incoherence enhancement.

Magnetic resonance imaging (MRI) is considered a key modality for the future as it offers several advantages, including the use of non-ionizing radiation and having no known side effects on the human body, and has recently begun to serve as a key component of multi-modal neuroimaging. However, two major intrinsic problems exist: slow acquisition and intrusive acoustic noise. Parallel MRI (pMRI) techniques accelerate acquisition by reducing the duration and coverage of conventional gradient encoding. The under-sampled k-space data is detected with several receiver coils surrounding the object, using distinct spatial encoding information for each coil element to reconstruct the image. However, this scanning remains slow compared to typical clinical imaging (e.g. X-ray CT). Compressed Sensing (CS), a sampling theory based on random sub-sampling, has potential to further reduce the sampling used in pMRI, accelerating acquisition further. In this work, we propose a new CS SENSE pMRI reconstruction model promoting joint sparsity across channels and enhancing mutual incoherence to improve reconstruction accuracy from limited k-space data. For fast image reconstruction and fair comparisons, all reconstructions are computed with split-Bregman and variable splitting techniques. Numerical results show that, with the introduced methods, reconstruction performance can be crucially improved with limited amount of k-space data.

Non-convex compressed sensing CT reconstruction based on tensor discrete Fourier slice theorem.

X-ray computed tomography (CT) scanners provide clinical value through high resolution and fast imaging. However, achievement of higher signal-to-noise ratios generally requires emission of more X-rays, resulting in greater dose delivered to the body of the patient. This is of concern, as higher dose leads to greater risk of cancer, particularly for those exposed at a younger age. Therefore, it is desirable to achieve comparable scan quality while limiting X-ray dose. One means to achieve this compound goal is the use of compressed sensing (CS). A novel framework is presented to combine CS theory with X-ray CT. According to the tensor discrete Fourier slice theorem, the 1-D DFT of discrete Radon transform data is exactly mapped on a Cartesian 2-D DFT grid. The nonuniform random density sampling of Fourier coefficients is made feasible by uniformly sampling projection angles at random. Application of the non-convex CS model further reduces the sufficient number of measurements by enhancing sparsity. The numerical results show that, with limited projection data, the non-convex CS model significantly improves reconstruction performance over the convex model.