Design, implement, and optimize pattern recognition schemes for biomedical data sets and images based on the following research interests: novel non-linear texture features, feature selection, supervised machine learning, non-linear dimensionality reduction, image processing, signal processing, clustering, data mining, data visualization, and statistics. Develop quantitative computer-aided detection systems using topological texture and structure features and state-of-the art supervised learning techniques (partially industry funded). Use MATLAB, C/C++, parallel programming, supercomputing, WEKA. Supervise and mentor graduate and undergraduate students.

Developed complex structure analysis methods and medical image processing algorithms for bone images from different modalities (MRI, CT, and radiographs). Calculated morphological, geometrical and fractal features. Wrote pattern recognition, filter, segmentation, image enhancement and compression tools in MATLAB and C++ using ITK. Performed data mining and statistical analysis of image properties. Embedded algorithms in GUIs. Mentored graduate students on their research projects.

Quantitative image analysis of the observed large-scale galaxy distribution (Sloan Digital Sky Survey). Developed pattern recognition algorithms and image filters using IDL and C++. Correlated spatial image patterns (e.g. local geometry, morphology, density) with physical properties of galaxies. Worked with SQL database and Python.

Modified and used existing numerical simulations (FORTRAN) of the large-scale galaxy distribution. Developed image filter algorithms in IDL and C++. Analyzed geometrical and morphological features. Compared different numerical models with observations.