I am a Postdoctoral Associate in Carl E. Ravin Advanced Imaging Laboratories at Duke University Medical Center.

My main reserch focus is applications of machine learning algorithms in image based medical decision support systems. My recent research projects aim at improving case-based classifiers for reduction of false positive marks in mammograms. In order to do so, we apply modern computational intelligence algorithms such as genetic algorithms or particle swarm optimization as well as more classical ones such as k-means clustering or linear discriminant analysis to improve decision algorithms and optimize case bases of the available computer-aided decision systems. My other research interest is impact of training and testing conditions (such as number of examples, prevalence of positive examples as well as training method and training parameters) on the performance evaluation of classifiers (specifically neural networks).

Although recently medical applications have been my main focus, I have also worked on the theory of the underlying algorithms (sensitivity analysis method for neural network pruning) and applications in other fields (e.g., control in multi-agent systems).