SENIHA ESEN YUKSEL, Ph.D.
CSE Bldg. Rm 542, University of Florida, Gainesville, FL 32611 USA
Cell: 352 – 870 4993
Machine learning, computer vision, statistical analysis and applied math.
Doctor of Philosophy in Computer Sciences & Engineering
University of Florida, Gainesville, USA
Chair: Prof. Paul D. Gader
Dissertation: Context-based classification via data-dependent mixtures of logistic and hidden
Markov model classifiers.
Master of Science in Electrical and Computer Engineering
University of Louisville, Louisville, Kentucky, USA
Chair: Prof. Aly A. Farag
Bachelor of Science in Electrical and Electronics Engineering
1999 – 2003
Middle East Technical University, Ankara, Turkey
August 2011 – Present
Department of Material Science and Engineering, University of Florida, Gainesville, FL, USA.
Working in an interdisciplinary project the Material Science and Engineering Department
and the Computer Engineering Department at UF, to develop tools for hyper-spectral data
SVM, end-member detection, hyper-spectral data classification.
2007 – 2011
Computational Science & Intelligence Lab, University of Florida, Gainesville, FL, USA
Worked on context-based multi-class classification for landmine detection using ground
penetrating radar (GPR) and wide-band electro-magnetic Induction (WEMI) data.
Implemented a fusion algorithm of GPR and WEMI confidence values using
Hierarchical Mixture of Experts that significantly increased the classification rates
while providing insightful results.
classification and derived a lower bound for it to prevent over-training.
Landmine detection, fusion, Bayesian and variational models.
Developed a novel model, Mixture of Hidden Markov Models Experts (MHMME), for
context-based multi-class classification of sequential data.
Context-based classification, hidden Markov models, mixture of
experts, 2D shape recognition, MATLAB
Developed a novel model based on multiple instance learning for the context-based
classification of time-series data with ambiguous features.
Multiple instance learning, sampling, hidden Markov models.