Faculty Bibliography

This letter proposes to estimate low-rank matrices by formulating a convex optimization problem with nonconvex regularization. We employ parameterized nonconvex penalty functions to estimate the nonzero singular values more accurately than the nuclear norm. A closed-form solution for the global optimum of the proposed objective function (sum of data fidelity and the nonconvex regularizer) is also derived. The solution reduces to singular value thresholding method as a special case. The proposed method is demonstrated for image denoising.

OBJECTIVE: Concussion is a major public health problem and considerable efforts are focused on sideline-based diagnostic testing to guide return-to-play decision-making and clinical care. The King-Devick (K-D) test, a sensitive sideline performance measure for concussion detection, reveals slowed reading times in acutely concussed subjects, as compared to healthy controls; however, the normal behavior of eye movements during the task and deficits underlying the slowing have not been defined. METHODS: Twelve healthy control subjects underwent quantitative eye tracking during digitized K-D testing. RESULTS: The total K-D reading time was 51.24 (+/-9.7) seconds. A total of 145 saccades (+/-15) per subject were generated, with average peak velocity 299.5 degrees /s and average amplitude 8.2 degrees . The average inter-saccadic interval was 248.4ms. Task-specific horizontal and oblique saccades per subject numbered, respectively, 102 (+/-10) and 17 (+/-4). Subjects with the fewest saccades tended to blink more, resulting in a larger amount of missing data; whereas, subjects with the most saccades tended to make extra saccades during line transitions. CONCLUSIONS: Establishment of normal and objective ocular motor behavior during the K-D test is a critical first step towards defining the range of deficits underlying abnormal testing in concussion. Further, it sets the groundwork for exploration of K-D correlations with cognitive dysfunction and saccadic paradigms that may reflect specific neuroanatomic deficits in the concussed brain.

This paper describes an exponential transient excision algorithm (ETEA). In biomedical time series analysis, e.g., in vivo neural recording and electrocorticography (ECoG), some measurement artifacts take the form of piecewise exponential transients. The proposed method is formulated as an unconstrained convex optimization problem, regularized by smoothed l(1)-norm penalty function, which can be solved by majorization-minimization (MM) method. With a slight modification of the regularizer, ETEA can also suppress more irregular piecewise smooth artifacts, especially, ocular artifacts (OA) in electro-encephalography (EEG) data. Examples of synthetic signal, EEG data, and ECoG data are presented to illustrate the proposed algorithms. (C) 2015 Elsevier B.V. All rights reserved.

This paper proposes a parametric model for saccadic waveforms. The model has a small number of parameters, yet it effectively simulates a variety of physiologic saccade properties. In particular, the model reproduces the established relationship between peak saccadic angular velocity and saccadic amplitude (i.e., the saccadic main sequence). The proposed saccadic waveform model can be used in the evaluation and validation of methods for quantitative saccade analysis. For example, we use the proposed saccade model to evaluate four well-known saccade detection algorithms. The comparison indicates the most reliable algorithm is one by Nystrom et al. We further use the proposed saccade model to evaluate the standard technique used for the estimation of peak saccadic angular velocity. The evaluation shows the occurrence of systematic errors. We thus suggest that saccadic angular velocity values determined by the standard technique (low-pass differentiation) should be interpreted and used with caution.

We propose a novel formulation for relaxed analysis-based sparsity in multiple dictionaries as a general type of prior for images, and apply it for Bayesian estimation in image restoration problems. Our formulation of a l2-relaxed l0 pseudo-norm prior allows for an especially simple maximum a posteriori estimation iterative marginal optimization algorithm, whose convergence we prove. We achieve a significant speedup over the direct (static) solution by using dynamically evolving parameters through the estimation loop. As an added heuristic twist, we fix in advance the number of iterations, and then empirically optimize the involved parameters according to two performance benchmarks. The resulting constrained dynamic method is not just fast and effective, it is also highly robust and flexible. First, it is able to provide an outstanding tradeoff between computational load and performance, in visual and objective, mean square error and structural similarity terms, for a large variety of degradation tests, using the same set of parameter values for all tests. Second, the performance benchmark can be easily adapted to specific types of degradation, image classes, and even performance criteria. Third, it allows for using simultaneously several dictionaries with complementary features. This unique combination makes ours a highly practical deconvolution method.