Compression is a standard procedure for making convolutional neural networks (CNNs) adhere to some specific computing resource constraints. However, searching for a compressed architecture typically involves a series of time-consuming training/valida
tion experiments to determine a good compromise between network size and performance accuracy. To address this, we propose an image complexity-guided network compression technique for biomedical image segmentation. Given any resource constraints, our framework utilizes data complexity and network architecture to quickly estimate a compressed model which does not require network training. Specifically, we map the dataset complexity to the target network accuracy degradation caused by compression. Such mapping enables us to predict the final accuracy for different network sizes, based on the computed dataset complexity. Thus, one may choose a solution that meets both the network size and segmentation accuracy requirements. Finally, the mapping is used to determine the convolutional layer-wise multiplicative factor for generating a compressed network. We conduct experiments using 5 datasets, employing 3 commonly-used CNN architectures for biomedical image segmentation as representative networks. Our proposed framework is shown to be effective for generating compressed segmentation networks, retaining up to $approx 95%$ of the full-sized network segmentation accuracy, and at the same time, utilizing $approx 32x$ fewer network trainable weights (average reduction) of the full-sized networks.
Image registration plays an important role in medical image analysis. Conventional optimization based methods provide an accurate estimation due to the iterative process at the cost of expensive computation. Deep learning methods such as learn-to-map
are much faster but either iterative or coarse-to-fine approach is required to improve accuracy for handling large motions. In this work, we proposed to learn a registration optimizer via a multi-scale neural ODE model. The inference consists of iterative gradient updates similar to a conventional gradient descent optimizer but in a much faster way, because the neural ODE learns from the training data to adapt the gradient efficiently at each iteration. Furthermore, we proposed to learn a modal-independent similarity metric to address image appearance variations across different image contrasts. We performed evaluations through extensive experiments in the context of multi-contrast 3D MR images from both public and private data sources and demonstrate the superior performance of our proposed methods.
Multi-contrast MRI images provide complementary contrast information about the characteristics of anatomical structures and are commonly used in clinical practice. Recently, a multi-flip-angle (FA) and multi-echo GRE method (MULTIPLEX MRI) has been d
eveloped to simultaneously acquire multiple parametric images with just one single scan. However, it poses two challenges for MULTIPLEX to be used in the 3D high-resolution setting: a relatively long scan time and the huge amount of 3D multi-contrast data for reconstruction. Currently, no DL based method has been proposed for 3D MULTIPLEX data reconstruction. We propose a deep learning framework for undersampled 3D MRI data reconstruction and apply it to MULTIPLEX MRI. The proposed deep learning method shows good performance in image quality and reconstruction time.
Retrospectively gated cine (retro-cine) MRI is the clinical standard for cardiac functional analysis. Deep learning (DL) based methods have been proposed for the reconstruction of highly undersampled MRI data and show superior image quality and magni
tude faster reconstruction time than CS-based methods. Nevertheless, it remains unclear whether DL reconstruction is suitable for cardiac function analysis. To address this question, in this study we evaluate and compare the cardiac functional values (EDV, ESV and EF for LV and RV, respectively) obtained from highly accelerated MRI acquisition using DL based reconstruction algorithm (DL-cine) with values from CS-cine and conventional retro-cine. To the best of our knowledge, this is the first work to evaluate the cine MRI with deep learning reconstruction for cardiac function analysis and compare it with other conventional methods. The cardiac functional values obtained from cine MRI with deep learning reconstruction are consistent with values from clinical standard retro-cine MRI.
Supersaturation is the fundamental parameter driving crystal formation, yet its dynamics during the growth of colloidal nanocrystals (NCs) are poorly understood. Experimental characterization of supersaturation in colloidal syntheses has been difficu
lt, limiting insight into the phenomena underlying NC growth. Hence, despite significant interest in the topic, how many types of NCs grow remain unclear. Here, we develop a framework to quantitatively characterize supersaturation in situ throughout NC growth. Using this approach, we investigate the seed-mediated synthesis of colloidal Au nanocubes, revealing a triphasic sequence for the supersaturation dynamics: rapid monomer consumption, sustained supersaturation, and then gradual monomer depletion. These NCs undergo different shape evolutions in different phases of the supersaturation dynamics. As shown with the Au nanocubes, we can use the supersaturation profile to theoretically predict the growth profile of NCs. We then apply these insights to rationally modulate shape evolutions, decreasing the yield of impurity NCs. Our findings demonstrate that the supersaturation dynamics of NC growth can be more complex than previously understood. While this study focuses experimentally on Au NCs, our framework is facile and applicable to a broad range of NCs undergoing classical growth. Thus, our methodology facilitates deeper understanding of the phenomena governing nanoscale crystal growth and provides insight towards the rational design of NCs.
From diagnosing neovascular diseases to detecting white matter lesions, accurate tiny vessel segmentation in fundus images is critical. Promising results for accurate vessel segmentation have been known. However, their effectiveness in segmenting tin
y vessels is still limited. In this paper, we study retinal vessel segmentation by incorporating tiny vessel segmentation into our framework for the overall accurate vessel segmentation. To achieve this, we propose a new deep convolutional neural network (CNN) which divides vessel segmentation into two separate objectives. Specifically, we consider the overall accurate vessel segmentation and tiny vessel segmentation as two individual objectives. Then, by exploiting the objective-dependent (homoscedastic) uncertainty, we enable the network to learn both objectives simultaneously. Further, to improve the individual objectives, we propose: (a) a vessel weight map based auxiliary loss for enhancing tiny vessel connectivity (i.e., improving tiny vessel segmentation), and (b) an enhanced encoder-decoder architecture for improved localization (i.e., for accurate vessel segmentation). Using 3 public retinal vessel segmentation datasets (CHASE_DB1, DRIVE, and STARE), we verify the superiority of our proposed framework in segmenting tiny vessels (8.3% average improvement in sensitivity) while achieving better area under the receiver operating characteristic curve (AUC) compared to state-of-the-art methods.
Quantum computing can become scalable through error correction, but logical error rates only decrease with system size when physical errors are sufficiently uncorrelated. During computation, unused high energy levels of the qubits can become excited,
creating leakage states that are long-lived and mobile. Particularly for superconducting transmon qubits, this leakage opens a path to errors that are correlated in space and time. Here, we report a reset protocol that returns a qubit to the ground state from all relevant higher level states. We test its performance with the bit-flip stabilizer code, a simplified version of the surface code for quantum error correction. We investigate the accumulation and dynamics of leakage during error correction. Using this protocol, we find lower rates of logical errors and an improved scaling and stability of error suppression with increasing qubit number. This demonstration provides a key step on the path towards scalable quantum computing.
Magnrtic flux ropes (MFRs) play a crucial role during magnetic reconnection. They are believed to be primarily generated by tearing mode instabilities in the electron diffusion region (EDR). However, they have never been observed inside the EDR. Here
, we present the first observation of an MFR inside an EDR. The bifurcated non-force-free MFR, with a width of 27.5de in the L direction and 4.8de in the N direction, was moving away from the X-line. Inside the MFR, strong energy dissipation was detected. The MFR can modulate the electric field in the EDR. We reconstructed magnetic topology of the electron-scale MFR. Our study promotes understanding of MRFs initial state and its role in electron-scale processes during magnetic reconnection.
Although the energies associated with nuclear reactions are due primarily to interactions involving nuclear forces, the rates and probabilities associated with those reactions are effectively governed by electromagnetic forces. Charges in the local e
nvironment can modulate the Coulomb barrier, and thereby change the rates of nuclear processes. Experiments are presented in which low-temperature electrons are attached to high-density rotating neutrals to form negative ions. The steady-state quiescent rotating plasma generated here lends itself to first prove the principle that low temperature systems can yield MeV fusion particles. It allows the use of high density of neutrals interacting with the wall to yield gain greater than unity. It also demonstrates that instabilities can be avoided with high neutral densities. Collective dynamic interactions within this steady-state quiescent plasma result in an arrangement of negative charges that lowers the effective Coulomb barrier to nuclear reactions at a solid wall of reactants. MeV alpha particles are synchronously observed with externally imposed pulses as evidence of fusion being enabled by Coulomb fields. Impacts on fusion, the source of energy in the universe, will be discussed.
We demonstrate spin-orbit torque (SOT) switching of amorphous CoTb single layer films with perpendicular magnetic anisotropy (PMA). The switching sustains even the film thickness is above 10 nm, where the critical switching current density keeps almo
st constant. Without the need of overcoming the strong interfacial Dzyaloshinskii-Moriya interaction caused by the heavy metal, a quite low assistant field of ~20 Oe is sufficient to realize the fully switching. The SOT effective field decreases and undergoes a sign change with the decrease of the Tb-concentration, implying that a combination of the spin Hall effect from both Co and Tb as well as an asymmetric spin current absorption accounts for the SOT switching mechanism. Our findings would advance the use of magnetic materials with bulk PMA for energy-efficient and thermal-stable non-volatile memories, and add a different dimension for understanding the ordering and asymmetry in amorphous thin films.
