PubMed Health⌕ Search

Biomedical subjects

Deniz Erdogmus

Publications and source records attributed to Deniz Erdogmus.

12 recordsLinked to original sources
PubMed2006

Method comparison of ultrasound and kilovoltage x-ray fiducial marker imaging for prostate radiotherapy targeting.

Several measurement techniques have been developed to address the capability for target volume reduction via target localization in image-guided radiotherapy; among these have been ultrasound (US) and fiducial marker (FM) software-assisted localization. In order to assess interchangeability between methods, US and FM localization were compared using established techniques for determination of agreement between measurement methods when a 'gold-standard' comparator does not exist, after performing both techniques daily on a sequential series of patients. At least 3 days prior to CT simulation, four gold seeds were placed within the prostate. FM software-assisted localization utilized the ExacTrac X-Ray 6D (BrainLab AG, Germany) kVp x-ray image acquisition system to determine prostate position; US prostate targeting was performed on each patient using the SonArray (Varian, Palo Alto, CA). Patients were aligned daily using laser alignment of skin marks. Directional shifts were then calculated by each respective system in the X, Y and Z dimensions before each daily treatment fraction, previous to any treatment or couch adjustment, as well as a composite vector of displacement. Directional shift agreement in each axis was compared using Altman-Bland limits of agreement, Lin's concordance coefficient with Partik's grading schema, and Deming orthogonal bias-weighted correlation methodology. 1,019 software-assisted shifts were suggested by US and FM in 39 patients. The 95% limits of agreement in X, Y and Z axes were +/-9.4 mm, +/-11.3 mm and +/-13.4, respectively. Three-dimensionally, measurements agreed within 13.4 mm in 95% of all paired measures. In all axes, concordance was graded as 'poor' or 'unacceptable'. Deming regression detected proportional bias in both directional axes and three-dimensional vectors. Our data suggest substantial differences between US and FM image-guided measures and subsequent suggested directional shifts. Analysis reveals that the vast majority of all individual US and FM directional measures may be expected to agree with each other within a range of 1-1.5 cm. Since neither system represents a gold standard, clinical judgment must dictate whether such a difference is of import. As IMRT protocols seek dose escalation and PTV reduction predicated on US- and FM-guided imaging, future studies are needed to address these potential clinically relevant issues regarding the interchangeability and accuracy of novel positional verification techniques. Comparison series with multiple image-guidance systems are needed to refine comparisons between targeting methods. However, we do not advocate interchangeability of US and FM localization methods.

Calibration↗
PubMed2006

Modeling and inverse controller design for an unmanned aerial vehicle based on the self-organizing map.

The next generation of aircraft will have dynamics that vary considerably over the operating regime. A single controller will have difficulty to meet the design specifications. In this paper, a self-organizing map (SOM)-based local linear modeling scheme of an unmanned aerial vehicle (UAV) is developed to design a set of inverse controllers. The SOM selects the operating regime depending only on the embedded output space information and avoids normalization of the input data. Each local linear model is associated with a linear controller, which is easy to design. Switching of the controllers is done synchronously with the active local linear model that tracks the different operating conditions. The proposed multiple modeling and control strategy has been successfully tested in a simulator that models the LoFLYTE UAV.

Aircraft↗
PubMed2006

Feature extraction using information-theoretic learning.

A classification system typically consists of both a feature extractor (preprocessor) and a classifier. These two components can be trained either independently or simultaneously. The former option has an implementation advantage since the extractor need only be trained once for use with any classifier, whereas the latter has an advantage since it can be used to minimize classification error directly. Certain criteria, such as Minimum Classification Error, are better suited for simultaneous training, whereas other criteria, such as Mutual Information, are amenable for training the feature extractor either independently or simultaneously. Herein, an information-theoretic criterion is introduced and is evaluated for training the extractor independently of the classifier. The proposed method uses nonparametric estimation of Renyi's entropy to train the extractor by maximizing an approximation of the mutual information between the class labels and the output of the feature extractor. The evaluations show that the proposed method, even though it uses independent training, performs at least as well as three feature extraction methods that train the extractor and classifier simultaneously.

Algorithms↗
PubMed2005

Interpreting spatial and temporal neural activity through a recurrent neural network brain-machine interface.

We propose the use of optimized brain-machine interface (BMI) models for interpreting the spatial and temporal neural activity generated in motor tasks. In this study, a nonlinear dynamical neural network is trained to predict the hand position of primates from neural recordings in a reaching task paradigm. We first develop a method to reveal the role attributed by the model to the sampled motor, premotor, and parietal cortices in generating hand movements. Next, using the trained model weights, we derive a temporal sensitivity measure to asses how the model utilized the sampled cortices and neurons in real-time during BMI testing.

Algorithms↗
PubMed2004

Image construction methods for phased array magnetic resonance imaging.

PURPOSE: To study image construction in phased array magnetic resonance imaging (MRI) systems from a statistical signal processing point of view. MATERIALS AND METHODS: Three new approaches for image combination with multiple coils are proposed: 1) one based on the singular value decomposition of the measurement matrix, which is asymptotically optimal in the signal-to-noise ratio sense; 2) one based on a maximum-likelihood formulation, incorporating a priori information on the coil sensitivities in a Bayesian manner; and 3) one based on a least-squares formulation, which incorporates a smoothness constraint on the coil sensitivities. RESULTS: Numerical examples using synthetic and real data are presented to illustrate the performance of these new approaches. Results on the synthetic data show improvement in signal-to-error ratio, while results on the real data (a 4.7 T four-coil image of a cat spinal cord) show that the proposed methods can improve the SNR in the final image by up to 3 dB in the regions of interest compared to conventional sum-of-squares processing. CONCLUSION: It is demonstrated that phased array MRI reconstruction performance can be improved by the use of more elaborate statistical signal processing algorithms.

Algorithms↗
PubMed2004

Feature selection in MLPs and SVMs based on maximum output information.

This paper presents feature selection algorithms for multilayer perceptrons (MLPs) and multiclass support vector machines (SVMs), using mutual information between class labels and classifier outputs, as an objective function. This objective function involves inexpensive computation of information measures only on discrete variables; provides immunity to prior class probabilities; and brackets the probability of error of the classifier. The maximum output information (MOI) algorithms employ this function for feature subset selection by greedy elimination and directed search. The output of the MOI algorithms is a feature subset of user-defined size and an associated trained classifier (MLP/SVM). These algorithms compare favorably with a number of other methods in terms of performance on various artificial and real-world data sets.

Algorithms↗
PubMed2004

Minimax mutual information approach for independent component analysis.

Minimum output mutual information is regarded as a natural criterion for independent component analysis (ICA) and is used as the performance measure in many ICA algorithms. Two common approaches in information-theoretic ICA algorithms are minimum mutual information and maximum output entropy approaches. In the former approach, we substitute some form of probability density function (pdf) estimate into the mutual information expression, and in the latter we incorporate the source pdf assumption in the algorithm through the use of nonlinearities matched to the corresponding cumulative density functions (cdf). Alternative solutions to ICA use higher-order cumulant-based optimization criteria, which are related to either one of these approaches through truncated series approximations for densities. In this article, we propose a new ICA algorithm motivated by the maximum entropy principle (for estimating signal distributions). The optimality criterion is the minimum output mutual information, where the estimated pdfs are from the exponential family and are approximate solutions to a constrained entropy maximization problem. This approach yields an upper bound for the actual mutual information of the output signals - hence, the name minimax mutual information ICA algorithm. In addition, we demonstrate that for a specific selection of the constraint functions in the maximum entropy density estimation procedure, the algorithm relates strongly to ICA methods using higher-order cumulants.

Algorithms↗
PubMed2003

SNR-optimality of sum-of-squares reconstruction for phased-array magnetic resonance imaging.

We consider the commonly used "Sum-of-Squares" (SoS) reconstruction method for phased-array magnetic resonance imaging with unknown coil sensitivities. We show that the signal-to-noise ratio (SNR) in the image produced by SoS is asymptotically (as the input SNR--> infinity ) equal to that of maximum-ratio combining, which is the best unbiased reconstruction method when the coil sensitivities are known. Finally, we discuss the implications of this result.

Computer Simulation↗
PubMedDate not supplied

A new classifier based on information theoretic learning with unlabeled data.

Supervised learning is conventionally performed with pairwise input-output labeled data. After the training procedure, the adaptive system's weights are fixed while the testing procedure with unlabeled data is performed. Recently, in an attempt to improve classification performance unlabeled data has been exploited in the machine learning community. In this paper, we present an information theoretic learning (ITL) approach based on density divergence minimization to obtain an extended training algorithm using unlabeled data during the testing. The method uses a boosting-like algorithm with an ITL based cost function. Preliminary simulations suggest that the method has the potential to improve the performance of classifiers in the application phase.

Algorithms↗
PubMedDate not supplied

Divide-and-conquer approach for brain machine interfaces: nonlinear mixture of competitive linear models.

This paper proposes a divide-and-conquer strategy for designing brain machine interfaces. A nonlinear combination of competitively trained local linear models (experts) is used to identify the mapping from neuronal activity in cortical areas associated with arm movement to the hand position of a primate. The proposed architecture and the training algorithm are described in detail and numerical performance comparisons with alternative linear and nonlinear modeling approaches, including time-delay neural networks and recursive multilayer perceptrons, are presented. This new strategy allows training the local linear models using normalized LMS and using a relatively smaller nonlinear network to efficiently combine the predictions of the linear experts. This leads to savings in computational requirements, while the performance is still similar to a large fully nonlinear network.

Artificial Intelligence↗
PubMedDate not supplied

Stochastic error whitening algorithm for linear filter estimation with noisy data.

Mean squared error (MSE) has been the most widely used tool to solve the linear filter estimation or system identification problem. However, MSE gives biased results when the input signals are noisy. This paper presents a novel stochastic gradient algorithm based on the recently proposed error whitening criterion (EWC) to tackle the problem of linear filter estimation in the presence of additive white disturbances. We will briefly motivate the theory behind the new criterion and derive an online stochastic gradient algorithm. Convergence proof of the stochastic gradient algorithm is derived making mild assumptions. Further, we will propose some extensions to the stochastic gradient algorithm to ensure faster, step-size independent convergence. We will perform extensive simulations and compare the results with MSE as well as total-least squares in a parameter estimation problem. The stochastic EWC algorithm has many potential applications. We will use this in designing robust inverse controllers with noisy data.

Algorithms↗