|
Work Experience:
|
2001-present: Research Scientist at Center for Excellence in Document
Analysis and Recognition (CEDAR) and Center for Unified Biometrics and Sensors (CUBS), University at Buffalo. Job includes research projects on character and word recognition, improving mail address reading
program for USPS and United Kingdom, research on automatic fingerprint identification and face recognition. Additional duties include publishing research results, writing grant proposals,
directing graduate student's research and reviewing submissions to scientific conferences and journals.
1999-2001: Research Assistant at Center for Excellence in Document
Analysis and Recognition (CEDAR), University at Buffalo. Job includes projects on
digit, zip code and word recognition. The job requires reading and understanding of existing code, extensive testing and
debugging, deep knowledge of pattern recognition algorithms.
May-August, 1998: Research Assistant at Center for Excellence in Document
Analysis and Recognition (CEDAR), University at Buffalo. Job included documenting
and testing of handwritten word recognition program.
1995-1999: Teaching assistant at the Department of Mathematics, University at Buffalo.
1992-1994: Part time job as programmer at Institute of System Research,
Moscow. Job included development of a program for estimating roots
of complex polynomials.
|
Education:
|
2000-2006: PhD program in the Department of Computer Science
and Engineering, University at Buffalo. Dissertation topic: "A Complexity Framework for Combination of Classifiers in Verification and Identification Systems".
Defense date: March 2006.
1998-2000: M.Sc. program at the Department of Computer Science
and Engineering, University at Buffalo. Degree received in May 2000.
1995-2000: Ph.D. program (ABD) at the Department of Mathematics, University at Buffalo. GPA: 3.875.
1993-1995: Graduate program in Mathematics in Moscow State University.
Passed qualifying examinations for the degree of candidate of physical
and mathematical sciences.
1988-1993: Moscow State University, Department of Mechanics and
Mathematics. Degree with honors in pure and applied mathematics.
Thesis: Invariant means on locally compact amenable groups.
|
Computer skills: |
Languages: C++/C, Matlab, Lisp, Java, Perl.
OS: Windows, UNIX (Solaris 2.6), Linux.
Other: ODBC, MFC, wxWidgets, SQL, YACC/LEX, Latex.
Strong mathematical background, pattern recognition and computer vision algorithms.
|
Projects: |
Facial expression biometrics. This project explores the possibility of identifying people by the way they express their emotions. The experiments confirm that geometric facial expression features can indeed help in identifying persons.
Fingerprint indexing system. This project explores the possibility of indexing fingerprints using minutia positions. In contrast to previous minutia indexing approaches, we construct a true index tree structure using local minutia neighborhood information. The system is expected to deliver a superior fingerprint matching time in identification systems.
Classifier combination in biometric identification and handwritten word recognition systems. I developed a general framework for classifier combination defining 4 types of combinations. Theory and experimental results show that combinations of most frequently used combination type provide only suboptimal performance. This project constitutes major part of my Ph.D. thesis research.
Secure fingerprint matching. I developed a unique algorithm for hashing fingerprint minutia information and matching hashed fingerprints. Only hash information is stored in the database and it is impossible to recover original fingerprints. Algorithm has good performance, and more effort is being applied to improve and expand it.
Barcode printing and recognition tools. Implemented 4-state barcode printing and recognition capabilities, and multiple interface changes.
Automatic fingerprint identification system. The system is a three layer client-server application.
Developed and implemented access to multiple databases through ODBC interface, most of the
server functions, some communication (DCOM) and client functions. Participated in the design
of the whole system and database. The system is currently being deployed in some US locations.
Barcode evaluation project. I supervised the project and implemented most of the software
for this project, including implementation of CRC, Reed-Solomon and convolutional encoders and
decoders. The results serve as a basis for barcode standard adoption by USPS.
United Kingdom Address Interpretation (UKAI) project. Improved handwritten word and phrase
recognition by training and combining word recognizers, and by implementing additional features.
Implemented character segmentation module for postal code.
Handwritten Address Interpretation (HWAI) project. Developed multiple
improvements to the handwritten word recognition module. Also developed efficient address block location module based on the contour clustering.
|
Patent: |
"Secure fingerprint matching by hashing localized information." Patent US 8005277.
|
Selected Publications: |
Tulyakov, S., C. Wu, and V. Govindaraju, On the Difference Between Optimal Combination Functions for Verification and Identification Systems. International Journal of Pattern Recognition and Artificial Intelligence, 2010. 24(2): p. 173-191.
Mansukhani, P., S. Tulyakov, and V. Govindaraju, A Framework for Efficient Fingerprint Identification Using a Minutiae Tree. Systems Journal, IEEE, 2010. 4(2): p. 126-137.
Tulyakov, S. and V. Govindaraju. Predicting Performance in Large-Scale Identification Systems by Score Resampling. in International Biometric Performance Conference (IBPC 2010). 2010.
Tulyakov, S. and V. Govindaraju. Neural Network Optimization for Combinations in Identification Systems. in Multiple Classifier Systems, 8th International Workshop, MCS 2009. 2009. Reykjavik, Iceland: Springer.
Tulyakov, S. and V. Govindaraju, Use of Identification Trial Statistics for Combination of Biometric Matchers. Information Forensics and Security, IEEE Transactions on, 2008. 3(4): p. 719-733.
Tulyakov, S. and V. Govindaraju, Issues and Advances in Biometrics, in Information Assurance, Security and Privacy Services, H.R. Rao and S.Upadhyaya, Editors. 2009, Emerald.
Tulyakov, S., F. Farooq, P. Mansukhani, and V. Govindaraju, Symmetric hash functions for secure fingerprint biometric systems. Pattern Recognition Letters, 2007. 28(16): p. 2427-2436.
Tulyakov, S., T. Slowe, Z. Zhang, and V. Govindaraju. Facial Expression Biometrics Using Tracker Displacement Features. in CVPR 2007 Workshop on Biometrics. 2007.
Tulyakov, S. and V. Govindaraju, Learning matching score dependencies for classifier combination, in Learning in Document Analysis and Recognition, S. Marinai and H. Fujisawa, Editors. 2007, Springer.
|