Browsing by Author "Zhang, Chi"
Now showing 1 - 3 of 3
Results Per Page
Sort Options
Item Deep Learning Approaches for Accelerated MRI Reconstruction(2023-05) Zhang, ChiAccelerated MRI is widely used clinically to reduce lengthy scan times, where datasets are sampled in Fourier domain (k-space) below Nyquist rate, and recovered using reconstruction methods that utilize redundancies in datasets or measurement hardware. MRI reconstruction methods either perform interpolation in k-space, or solve an inverse problem with a known forward operator to recover the desired image. Recently, deep learning (DL) has been investigated for both types of MRI reconstruction, with improved image quality at higher acceleration rates. Among k-space interpolation methods, robust artificial-neural-networks for k-space interpolation (RAKI) has been proposed to perform nonlinear k-space interpolation, showing improved noise resilience compared to its linear predecessor (GRAPPA). However, RAKI is time-consuming since it trains multiple convolutional neural networks (CNNs) to reconstruct a single acquisition, and RAKI shows blurring at high acceleration rates, both hindering its applicability in practice. For the inverse problem type methods, physics-guided deep learning (PG-DL) has been shown to offer improved image quality and robustness than purely data-driven approaches. In PG-DL, a regularized inverse problem is solved alternatively between a proximal step solved implicitly by a neural network and a linear data fidelity (DF) step for several unrolled iterations. As a sufficient number of iterations is necessary for both unrolling and internally for DF, PG-DL leads to a deep computational graph that consumes considerable GPU memory during training. Consequently it is challenging to apply PG-DL with several acquisition types, include high-resolution, 3D/4D, and non-Cartesian MRI. Furthermore, for any DL-based approach, recent studies for single-coil datasets suggest they exhibit instabilities against adversarial perturbations, which lead no visible differences at input but impact the reconstruction results. However, the impact of adversarial perturbations against conventional multi-coil MRI remains uninvestigated, which is the more clinically used acquisition setup. This thesis studies DL-based approaches for MRI reconstruction in terms of their practical translation. First, we investigate scan-specific DL methods, starting with strategies to reduce the processing time of RAKI from hours to a clinically acceptable range within seconds. This is achieved by utilizing multi-processing and a novel line-by-line network architecture which reduces the number of CNNs trained during reconstruction. We further propose residual RAKI (rRAKI) that performs hybrid linear and nonlinear k-space interpolation, where a linear convolution provides baseline reconstruction jointly with a nonlinear component for noise and artifact reduction. Compared to linear methods, rRAKI exhibits improved noise resilience, similar to RAKI, while rRAKI offers improved image sharpness compared to RAKI. We then move to database DL approaches, focusing on PG-DL. In this setting, we tackle the memory consumption issues for large-scale PG-DL. By combining several memory-efficient techniques, PG-DL of 3D, high-resolution, multi-coil, non-Cartesian MRI is achieved. These techniques include check-pointing that keeps only one unrolled step on GPU; a Toeplitz characterization for non-Cartesian encoding and decoding operations that replace memory consuming convolutional operations with point-wise multiplications; distributed learning techniques that exploit linearity of multi-coil DF terms to distribute it into multiple GPUs; and mixed precision training that trains image-domain CNN via half-precision, and DF that works with k-space data in single precision. We then propose a 2.5D PG-DL for 3D reconstruction to tackle the issue of insufficient training data in these 3D non-Cartesian applications. The 2.5D PG-DL consists of three 2D CNNs that treat 3D volumes as batches of 2D images from three orthogonal views, such that the 2D CNNs have access to a sufficient number of 2D images from limited 3D volumes during the training. The 3D reconstruction is formulated using three constraints over 2D views and finally unrolled via variable splitting with quadratic penalty. Our results suggest that the proposed 2.5D PG-DL offer improved image sharpness compared to conventional 3D PG-DL for limited training data. Finally, we investigate the impact of adversarial attacks on conventional multi-coil MRI techniques including SENSE, GRAPPA and compressed sensing. Our results show that adversarial perturbations generated on the zero-filled image, with constraints on its l∞ norm, impact all conventional multi-coil MRI techniques, leading to substantial degradation in reconstruction at high acceleration rates. Our results also suggest that the main cause of failure in PG-DL methods against adversarial attack is the linear DF step rather than the nonlinear regularizer.Item Supporting data for Temperature-dependent thermal conductivity of MBE-grown epitaxial SrSnO₃ films(2023-11-06) Zhang, Chi; Liu, Fengdeng; Guo, Silu; Zhang, Yingying; Xu, Xiaotian; Mkhoyan, Andre; Jalan, Bharat; Wang, Xiaojia; wang4940@umn.edu; Wang, Xiaojia; Materials Research Science & Engineering CenterThis work studies the temperature-dependent thermal properties of a single crystalline SSO thin film prepared with hybrid molecular beam epitaxy. By combining time-domain thermoreflectance and Debye–Callaway modeling, physical insight into thermal transport mechanisms is provided. At room temperature, the 350-nm SSO film has a thermal conductivity of 4.4 W m¯¹ K¯¹ , ∼60% lower than those of other perovskite oxides (SrTiO₃, BaSnO₃) with the same ABO₃ structural formula. This difference is attributed to the low zone-boundary frequency of SSO, resulting from its distorted orthorhombic structure with tilted octahedra. At high temperatures, the thermal conductivity of SSO decreases with temperature following a ∼T ¯⁰∙⁵⁴ dependence, weaker than the typical T¯¹ trend dominated by the Umklapp scattering. Corresponding author for STEM data is K. Andre Mkhoyan. Corresponding author for film growth and XRD data is Bharat Jalan. Corresponding author for TDTR data is Xiaojia Wang.Item Supporting data for Wide-range continuous tuning of the thermal conductivity of La0.5Sr0.5CoO3−δ films via room-temperature ion-gel gating(2023-04-19) Zhang, Yingying; Postiglione, William M; Xie, Rui; Zhang, Chi; Zhou, Hao; Chaturvedi, Vipul; Heltemes, Kei; Zhou, Hua; Feng, Tianli; Leighton, Chris; Wang, Xiaojia; wang4940@umn.edu; Wang, Xiaojia; Materials Research Science & Engineering CenterThese files contain data along with associated output from instrumentation supporting all results reported in Wide-range continuous tuning of the thermal conductivity of La0.5Sr0.5CoO3-delta films via room-temperature ion-gel gating. Solid-state control of the thermal conductivity of materials is of exceptional interest for novel devices such as thermal diodes and switches. Here, we demonstrate the ability to continuously tune the thermal conductivity of nanoscale films of La0.5Sr0.5CoO3-delta (LSCO) by a factor of over 5, via a room-temperature electrolyte-gate-induced non-volatile topotactic phase transformation from perovskite (with ≈ 0.1) to an oxygen-vacancy-ordered brownmillerite phase (with = 0.5), accompanied by a metal-insulator transition. Combining time-domain thermoreflectance and electronic transport measurements, model analyses based on molecular dynamics and Boltzmann transport, and structural characterization by X-ray diffraction, we uncover and deconvolve the effects of these transitions on heat carriers, including electrons and lattice vibrations. The wide-range continuous tunability of LSCO thermal conductivity enabled by low-voltage (below 4 V) room-temperature electrolyte gating opens the door to non-volatile dynamic control of thermal transport in perovskite-based functional materials, for thermal regulation and management in device applications. Authors to whom correspondence should be addressed are Chris Leighton (leighton@umn.edu) and Xiaojia Wang (wang4940@umn.edu).