Physical Sciences and Mathematics Commons^™

Automatic Hemorrhage Segmentation In Brain Ct Scans Using Curriculum-Based Semi-Supervised Learning, Solayman H. Emon, Tzu-Liang (Bill) Tseng, Michael Pokojovy, Peter Mccaffrey, Scott Moen, Md Fashiar Rahman

Mathematics & Statistics Faculty Publications

One of the major neuropathological consequences of traumatic brain injury (TBI) is intracranial hemorrhage (ICH), which requires swift diagnosis to avert perilous outcomes. We present a new automatic hemorrhage segmentation technique via curriculum-based semi-supervised learning. It employs a pre-trained lightweight encoder-decoder framework (MobileNetV2) on labeled and unlabeled data. The model integrates consistency regularization for improved generalization, offering steady predictions from original and augmented versions of unlabeled data. The training procedure employs curriculum learning to progressively train the model at diverse complexity levels. We utilize the PhysioNet dataset to train and evaluate the proposed approach. The performance results surpass those of …

Infusing Machine Learning And Computational Linguistics Into Clinical Notes, Funke V. Alabi, Onyeka Omose, Omotomilola Jegede

Mathematics & Statistics Faculty Publications

Entering free-form text notes into Electronic Health Records (EHR) systems takes a lot of time from clinicians. A large portion of this paper work is viewed as a burden, which cuts into the amount of time doctors spend with patients and increases the risk of burnout. We will see how machine learning and computational linguistics can be infused in the processing of taking clinical notes. We are presenting a new language modeling task that predicts the content of notes conditioned on historical data from a patient's medical record, such as patient demographics, lab results, medications, and previous notes, with the …

Sparse Representer Theorems For Learning In Reproducing Kernel Banach Spaces, Rui Wang, Yuesheng Xu, Mingsong Yan

Mathematics & Statistics Faculty Publications

Sparsity of a learning solution is a desirable feature in machine learning. Certain reproducing kernel Banach spaces (RKBSs) are appropriate hypothesis spaces for sparse learning methods. The goal of this paper is to understand what kind of RKBSs can promote sparsity for learning solutions. We consider two typical learning models in an RKBS: the minimum norm interpolation (MNI) problem and the regularization problem. We first establish an explicit representer theorem for solutions of these problems, which represents the extreme points of the solution set by a linear combination of the extreme points of the subdifferential set, of the norm function, …

Generalized Sparse Bayesian Learning And Application To Image Reconstruction, Jan Glaubitz, Anne Gelb, Guohui Song

Mathematics & Statistics Faculty Publications

Image reconstruction based on indirect, noisy, or incomplete data remains an important yet challenging task. While methods such as compressive sensing have demonstrated high-resolution image recovery in various settings, there remain issues of robustness due to parameter tuning. Moreover, since the recovery is limited to a point estimate, it is impossible to quantify the uncertainty, which is often desirable. Due to these inherent limitations, a sparse Bayesian learning approach is sometimes adopted to recover a posterior distribution of the unknown. Sparse Bayesian learning assumes that some linear transformation of the unknown is sparse. However, most of the methods developed are …

Fast Multiscale Functional Estimation In Optimal Emg Placement For Robotic Prosthesis Controllers, Jin Ren, Guohui Song, Lucia Tabacu, Yuesheng Xu

Mathematics & Statistics Faculty Publications

Electromyogram (EMG) signals play a significant role in decoding muscle contraction information for robotic hand prosthesis controllers. Widely applied decoders require a large amount of EMG signals sensors, resulting in complicated calculations and unsatisfactory predictions. By the biomechanical process of single degree-of-freedom human hand movements, only several EMG signals are essential for accurate predictions. Recently, a novel predictor of hand movements adopted a multistage sequential adaptive functional estimation (SAFE) method based on the historical functional linear model (FLM) to select important EMG signals and provide precise projections.

However, SAFE repeatedly performs matrix-vector multiplications with a dense representation matrix of the …

A Super Fast Algorithm For Estimating Sample Entropy, Weifeng Liu, Ying Jiang, Yuesheng Xu

Mathematics & Statistics Faculty Publications

: Sample entropy, an approximation of the Kolmogorov entropy, was proposed to characterize complexity of a time series, which is essentially defined as − log(B/A), where B denotes the number of matched template pairs with length m and A denotes the number of matched template pairs with m + 1, for a predetermined positive integer m. It has been widely used to analyze physiological signals. As computing sample entropy is time consuming, the box-assisted, bucket-assisted, x-sort, assisted sliding box, and kd-tree-based algorithms were proposed to accelerate its computation. These algorithms require O(N²) or …

Deeply Learning Deep Inelastic Scattering Kinematics, Markus Diefenthaler, Abdullah Farhat, Andrii Verbytskyi, Yuesheng Xu

Mathematics & Statistics Faculty Publications

We study the use of deep learning techniques to reconstruct the kinematics of the neutral current deep inelastic scattering (DIS) process in electron–proton collisions. In particular, we use simulated data from the ZEUS experiment at the HERA accelerator facility, and train deep neural networks to reconstruct the kinematic variables Q² and x. Our approach is based on the information used in the classical construction methods, the measurements of the scattered lepton, and the hadronic final state in the detector, but is enhanced through correlations and patterns revealed with the simulated data sets. We show that, with the appropriate selection …

Statistical Analysis And Comparison Of Optical Classification Of Atmospheric Aerosol Lidar Data, Mohammed Alqawba, Norou Diawara, Kwasi G. Afrifa, Mohamed I. Elbakary, Mecit Cetin, Khan Iftekharuddin

Mathematics & Statistics Faculty Publications

In this article, we present a new study for the analysis and classification of atmospheric aerosols in remote sensing LIDAR data. Information on particle size and associated properties are extracted from these remote sensing atmospheric data which are collected by a ground-based LIDAR system. This study first considers optical LIDAR parameter-based classification methods for clustering and classification of different types of harmful aerosol particles in the atmosphere. Since accurate methods for aerosol prediction behaviors are based upon observed data, computational approaches must overcome design limitations, and consider appropriate calibration and estimation accuracy. Consequently, two statistical methods based on generalized linear …

Multiplicative Noise Removal: Nonlocal Low-Rank Model And It's Proximal Alternating Reweighted Minimization Algorithm, Xiaoxia Liu, Jian Lu, Lixin Shen, Chen Xu, Yuesheng Xu

Mathematics & Statistics Faculty Publications

The goal of this paper is to develop a novel numerical method for efficient multiplicative noise removal. The nonlocal self-similarity of natural images implies that the matrices formed by their nonlocal similar patches are low-rank. By exploiting this low-rank prior with application to multiplicative noise removal, we propose a nonlocal low-rank model for this task and develop a proximal alternating reweighted minimization (PARM) algorithm to solve the optimization problem resulting from the model. Specifically, we utilize a generalized nonconvex surrogate of the rank function to regularize the patch matrices and develop a new nonlocal low-rank model, which is a nonconvex …

Sparsity Promoting Regularization For Effective Noise Suppression In Spect Image Reconstruction, Wei Zheng, Si Li, Andrzej Krol, C. Ross Schmidtlein, Xueying Zeng, Yuesheng Xu

Mathematics & Statistics Faculty Publications

The purpose of this research is to develop an advanced reconstruction method for low-count, hence high-noise, Single-Photon Emission Computed Tomography (SPECT) image reconstruction. It consists of a novel reconstruction model to suppress noise while conducting reconstruction and an efficient algorithm to solve the model. A novel regularizer is introduced as the nonconvex denoising term based on the approximate sparsity of the image under a geometric tight frame transform domain. The deblurring term is based on the negative log-likelihood of the SPECT data model. To solve the resulting nonconvex optimization problem a Preconditioned Fixed-point Proximity Algorithm (PFPA) is introduced. We prove …

Development Of The Electron Cooling Simulation Program For Jleic, H. Zhang, J. Chen, R. Li, Y. Zhang, H. Huang, L. Luo

Mathematics & Statistics Faculty Publications

In the JLab Electron Ion Collider (JLEIC) project the traditional electron cooling technique is used to reduce the ion beam emittance at the booster ring, and to compensate the intrabeam scattering effect and maintain the ion beam emittance during collision at the collider ring. A new electron cooling process simulation program has been developed to fulfill the requirements of the JLEIC electron cooler design. The new program allows the users to calculate the electron cooling rate and simulate the cooling process with either DC or bunched electron beam to cool either coasting or bunched ion beam. It has been benchmarked …

Lattice-Boltzmann Simulations Of The Thermally Driven 2d Square Cavity At High Rayleigh Numbers, Dario Contrino, Pierre Lallemand, Pietro Asinari, Li-Shi Luo

Mathematics & Statistics Faculty Publications

The thermal lattice Boltzmann equation (TLBE) with multiple-relaxation-times (MRT) collision model is used to simulate the steady thermal convective flows in the two-dimensional square cavity with differentially heated vertical walls at high Rayleigh numbers. The MRT-TLBE consists of two sets of distribution functions, i.e., a D2Q9 model for the mass-momentum equations and a D2Q5 model for the temperature equation. The dimensionless flow parameters are the following: the Prandtl number Pr = 0.71 and the Rayleigh number Ra = 10⁶, 10⁷, and 10⁸. The D2Q9 + D2Q5 MRT-TLBE is shown to be second-order accurate and …

Lattice Boltzmann Simulations Of Thermal Convective Flows In Two Dimensions, Jia Wang, Donghai Wang, Pierre Lallemand, Li-Shi Luo

Mathematics & Statistics Faculty Publications

In this paper we study the lattice Boltzmann equation (LBE) with multiple-relaxation-time (MRT) collision model for incompressible thermo-hydrodynamics with the Boussinesq approximation. We use the MRT thermal LBE (TLBE) to simulate the following two flows in two dimensions: the square cavity with differentially heated vertical walls and the Rayleigh-Benard convection in a rectangle heated from below. For the square cavity, the flow parameters in this study are the Rayleigh number Ra = 10³-10⁶, and the Prandtl number Pr = 0.71; and for the Rayleigh-Benard convection in a rectangle, Ra = 2 ^. 10³, 10 …

A Two-Population Insurgency In Colombia: Quasi-Predator-Prey Models - A Trend Towards Simplicity, John A. Adam, John A. Sokolowski, Catherine M. Banks

Mathematics & Statistics Faculty Publications

A sequence of analytic mathematical models has been developed in the context of the "low-level insurgency" in Colombia, from 1993 to the present. They are based on generalizations of the two-population "predator-prey" model commonly applied in ecological modeling, and interestingly, the less sophisticated models yield more insight into the problem than the more complicated ones, but the formalism is available to adapt the model "upwards" in the event that more data becomes available, or as the situation increases in complexity. Specifically, so-called "forcing terms" were included initially in the coupled differential equations to represent the effects of government policies towards …

A Simplified Model For Growth Factor Induced Healing Of Wounds, F. J. Vermolen, E. Van Baaren, J. A. Adam

Mathematics & Statistics Faculty Publications

A mathematical model is developed for the rate of healing of a circular or elliptic wound. In this paper the regeneration, decay and transport of a generic 'growth factor', which induces the healing of the wound, is taken into account. Further, an equation of motion is derived for radial healing of a circular wound. The expressions for the equation of motion and the distribution of the growth factor are related in such a way that no healing occurs if the growth factor concentration at the wound edge is below a threshold value. In this paper we investigate the influence of …

Protocols For Disease Classification From Mass Spectrometry Data, Michael Wagner, Dayanand Naik, Alex Pothen

Mathematics & Statistics Faculty Publications

We report our results in classifying protein matrix-assisted laser desorption/ionizationtime of flight mass spectra obtained from serum samples into diseased and healthy groups. We discuss in detail five of the steps in preprocessing the mass spectral data for biomarker discovery, as well as our criterion for choosing a small set of peaks for classifying the samples. Cross-validation studies with four selected proteins yielded misclassification rates in the 10-15% range for all the classification methods. Three of these proteins or protein fragments are down-regulated and one up-regulated in lung cancer, the disease under consideration in this data set. When cross-validation studies …

A Simplified Model Of Wound Healing - Ii: The Critical Size Defect In Two Dimensions, J. S. Arnold, John A. Adam

Mathematics & Statistics Faculty Publications

Recently, a one-dimensional model was developed which gives a reasonable explanation for the existence of a Critical Size Defect (CSD) in certain animals [1]. In this paper, we examine the more realistic two-dimensional model of a circular wound of uniform depth to see what modifications are to be found, as compared with the one-dimensional model, in studying the CSD phenomenon. It transpires that the range of CSD sizes for a reasonable estimate of parameter values is 1 mm-1 cm. More realistic estimates await the appropriate experimental data.

A Simplified Model Of Wound Healing (With Particular Reference To The Critical Size Defect), J. A. Adam

Mathematics & Statistics Faculty Publications

This paper is an attempt to construct a simple mathematical model of wound healing/tissue regeneration which reproduces some of the known qualitative features of those phenomena. It does not address the time development of the wound in any way, but does examine conditions (e.g., wound size) under which such healing may occur. Two related one-dimensional models are examined here. The first, and simpler of the two corresponds to a "swath" of tissue (or more realistically in this case, bone) removed from an infinite plane of tissue in which only a thin band of tissue at the wound edges takes part …

The Adjoint Alternative For Matrix Operators, C. H. Cooke

Mathematics & Statistics Faculty Publications

The following inverse problem is considered: given a matrix B of rank r, does there exist a matrix A such that

B = T(A) = adjoint (A)

where the classical adjoint operation is intended? Conditions are determined on the rank of B which decides whether or not B lies in the range of the matrix adjoint operator.

Corrigendum To “Post-Surgical Passive Response Of Local Environment To Primary Tumor Removal”: Mathl. Comput. Modelling, Vol. 25, No. 6, Pp. 7–17, 1997, J. A. Adam, C. Bellomo

Mathematics & Statistics Faculty Publications

The computer program that was used to generate the graphs for the concentration of inhibitor contained an error. This influenced the scaling in the original Figures 2 and 3. As an example, a sample of the corrected graphs are given below. Copies of other corrected figures can be obtained from the authors. It is important to note that the “pulse” appears for the function rC(r, t). As can be seen, it travels slowly outward with decreasing amplitude. The mathematical analysis in the paper remains unchanged.

Post-Surgical Passive Response Of Local Environment To Primary Tumor Removal, J. A. Adam, C. Bellomo

Mathematics & Statistics Faculty Publications

Prompted by recent clinical observations on the phenomenon of metastasis inhibition by an angiogenesis inhibitor, a mathematical model is developed to describe the post-surgical response of the local environment to the “surgical” removal of a spherical tumor in an infinite homogeneous domain. The primary tumor is postulated to be a source of growth inhibitor prior to its removal at t = 0; the resulting relaxation wave arriving from the disturbed (previously steady) state is studied, closed form analytic solutions are derived, and the asymptotic speed of the pulse is estimated to be about 2 × 10⁻⁴ cm/sec for the …

The Hadamard Matroid And An Anomaly In Its Single Element Extensions, C. H. Cooke

Mathematics & Statistics Faculty Publications

A nonstandard vector space is formulated, whose bases afford a representation of what is called a Hadamard matroid, M_p. For prime p, existence of M_p is equivalent to the existence of both a classical Hadamard matrix H(p,p) and a certain affine resolvable, balanced incomplete block design AR(p). An anomaly in the representable single element extension of a Hadamard matroid is discussed.

Antiplane Shear Of A Strip Containing A Staggered Array Of Rigid Line Inclusions, G. Kerr, G. Melrose, J. Tweed

Mathematics & Statistics Faculty Publications

Motivated by the increased use of fibre-reinforced materials, we illustrate how the effective elastic modulus of an isotropic and homogeneous material can be increased by the insertion of rigid inclusions. Specifically we consider the two-dimensional antiplane shear problem for a strip of material. The strip is reinforced by introducing two sets of ribbon-like, rigid inclusions perpendicular to the faces of the strip. The strip is then subjected to a prescribed uniform displacement difference between its faces, see Figure 1. it should be noted that the problem posed is equivalent to that of the uniform antiplane shear problem for an infinite …

An Efficient Runge-Kutta (4,5) Pair, P. Bogacki, L. F. Shampine

Mathematics & Statistics Faculty Publications

A pair of explicit Runge-Kutta formulas of orders 4 and 5 is derived. It is significantly more efficient than the Fehlberg and Dormand-Prince pairs, and by standard measures it is of at least as high quality. There are two independent estimates of the local error. The local error of the interpolant is, to leading order, a problem-independent function of the local error at the end of the step.

Data Compression Based On The Cubic B-Spline Wavelet With Uniform Two-Scale Relation, S. K. Yang, C. H. Cooke

Mathematics & Statistics Faculty Publications

The aim of this paper is to investigate the potential artificial compression which can be achieved using an interval multiresolution analysis based on a semiorthogonal cubic B-spline wavelet. The Chui-Quak [1] spline multiresolution analysis for the finite interval has been modified [2] so as to be characterized by natural spline projection and uniform two-scale relation. Strengths and weaknesses of the semiorthogonal wavelet as regards artificial compression and data smoothing by the method of thresholding wavelet coefficients are indicated.

A Family Of Parallel Runge-Kutta Pairs, P. Bogacki

Mathematics & Statistics Faculty Publications

Increasing availability of parallel computers has recently spurred a substantial amount of research concerned with designing explicit Runge-Kutta methods to be implemented on such computers. Here, we discuss a family of methods that require fewer processors than methods presently available do, still achieving a similar speed-up. In particular, (5,6) and (6,7) pairs are derived, that require a minimum number of function evaluations on two and three processors, respectively.

Error Estimates And Lipschitz Constants For Best Approximation In Continuous Function Spaces, M. Bartelt, W. Li

Mathematics & Statistics Faculty Publications

We use a structural characterization of the metric projection PG(f), from the continuous function space to its one-dimensional subspace G, to derive a lower bound of the Hausdorff strong unicity constant (or weak sharp minimum constant) for PG and then show this lower bound can be attained. Then the exact value of Lipschitz constant for PG is computed. The process is a quantitative analysis based on the Gâteaux derivative of PG, a representation of local Lipschitz constants, the equivalence of local and global Lipschitz constants for lower semicontinuous mappings, and construction …

A Mathematical Model Of Cycle Chemotherapy, J. C. Panetta, J. Adam

Mathematics & Statistics Faculty Publications

A mathematical model is used to discuss the effects of cycle-specific chemotherapy. The model includes a constraint equation which describes the effects of the drugs on sensitive normal tissue such as bone marrow. This model investigates both pulsed and piecewise-continuous chemotherapeutic effects and calculates the parameter regions of acceptable dose and period. It also identifies the optimal period needed for maximal tumor reduction. Examples are included concerning the use of growth factors and how they can enhance the cell kill of the chemotherapeutic drugs.

Temporal Model Of An Optically Pumped Co-Doped Solid State Laser, T. G. Wangler, J. J. Swetits, A. M. Buoncristiani

Mathematics & Statistics Faculty Publications

Currently, research is being conducted on the optical properties of materials associated with the development of solid-state lasers in the 2 micron region. In support of this effort, a mathematical model describing the energy transfer in a holmium laser sensitized with thulium is developed. In this paper, we establish some qualitative properties of the solution of the model, such as non-negativity, boundedness, and integrability. A local stability analysis is then performed from which conditions for asymptotic stability are obtained. Finally, we report on our numerical analysis of the system and how it compares with experimental results.

The Dynamics Of Growth-Factor-Modified Immune-Response To Cancer Growth: One-Dimensional Models, J. A. Adam

Mathematics & Statistics Faculty Publications

By characterizing the effect of tumor growth factors as deviations from normal logistic-type growth rates, the spatio-temporal dynamics for a one-dimensional model of cancer growth incorporating immune response are studied. The growth rates considered are classified respectively as normal, activated, inhibited and delay activated. The homogeneous steady states are defined by relative extrema of a ''free energy'' function V(x) for each of the above four cases. This function is of particular importance in studying the coexistence of tumoral and cancer-free steady states, and in identifying the nature (progressive or regressive) of travelling wave solutions to the nonlinear partial differential equation …