Pattern Recognition 

Machine Learning and Perception

Computer Vision

Image Processing

Adaptive Artificial Intelligence 

Data Mining and Knowledge Discovery

Discrete Mathematics.

Including the following topics:

Structural Hidden Markov Models (developed by Dr. Bouchaffra)

Topological Hidden Markov Models (developed by Dr. Bouchaffra)

Classification based on Markov Chains, Monte Carlo Methods

Partially Observed Markov Decision Processes

Structural Hidden Markov Decision Processes

Entropy Minimization

Rough Set Theory

 Game Theory

Hidden Markov Models

Stochastic Petri-Nets

Emergent Computation

Some Applications:

Biometrics

Speech and Handwriting Recognition

Medical Computer Vision

Control Systems and Robotics

 Computer Games 

Proteins Classification

Automatic Surveillance and Human Behavior Control

Conferences Involvement

General Chair and One of the Organizers of the:

• International Computer System and Information Technology, ICSIT'05 (IEEE/ CDTA)

          July 19-22, 2005 Hilton Hotel, Algiers, http://www.cdta.dz/manifestations/icsit05/index.htm

Program Committee Member:

• 2012 IEEE World Congress On Computational Intelligence (WCCI), June 10-15, International Convention Center, Brisbane, Australia, http://www-lipn.univ-paris13.fr/~grozavu/IJCNN/default.html

• Third International Conference on Agents and Artificial Intelligence (ICAART 2011), 28-30 January 2011, Rome, Italy http://www.icaart.org/

• International Conference on Machine and Web Intelligence, ICMWI'2010, October 3-5, Algiers, 2010, http://www.icmwi2010.org

• The Fifth International Workshop on Artificial Neural Networks and Intelligent Information Processing -ANNIIP 2009, 4-5 July, 2009, Milan, Italy http://www.icinco.org/ANNIIP.htm

• The 13th IASTED International Conference on Artificial Intelligence and Soft Computing ASC 2009 ~September 7-9, 2009, Palma de Mallorca, Spain http://www.iasted.org/conferences/home-683.html

• International Conference on Agents and Artificial Intelligence, ICAART'2009, 19-21 Januray 2009, Porto, Portugal, http://www.icaart.org/Program_Committee.htm

• The 19th International Conference On Pattern Recognition (ICPR), Dec. 8-11, 2008, Tampa Convention Center: Tampa, Florida, USA, http://www.icpr2008.org

• IADIS Multi-Conference on Computer Science and Information Systems: International Conference: Intelligent Systems and Agents, Amsterdam, Netherlands, 22-27 July 2008
  http://www.isa-conf.org/

• The first IASTED International Conference on Visual Communications (VC 2008), March 17 – 19, 2008, Innsbruck, Austria, http://iasted.org/conferences/home-613.html
  

• The 11th IASTED International Conference on Artificial Intelligence & Soft Computing ASC 2007 ~August 29-31, 2007, Palma de Mallorca, Spain http://www.iasted.org/conferences/cfp-584.html?ci

• IADIS Multi Conference on Computer Science and Intelligent Systems: International Conference: Intelligent Systems and Agents, Lisbon. Portugal, 3-5 July 2007, http://www.isa-conf.org/

• The 9th International Conference on Enterprise Information Systems, 12-16, June 2007, Funchal, Madeira - Portugal (http://www.iceis.org)

• The 10th IASTED International Conference on Artificial Intelligence & Soft Computing ASC 2006 ~August 28-30, 2006, Palma de Mallorca, Spain
  http://www.iasted.org/conferences/2006/spain/committee-544.htm

• The 9th IASTED International Conference on Artificial Intelligence & Soft Computing ASC September 12-14, 2005 Benidorm, Spain
  http://www.iasted.org/conferences/2005/spain/c481.htm

• The Irish Machine Vision and Image Processing Conference (IMVIP 2005) 
  August 30-September 1, 2005 Queen's University Belfast, Northern Ireland http://www.cs.qub.ac.uk/imvip2005

•   ITSC'05- 8th International IEEE Conference on Intelligent Transportation Systems,
   13-16 September, 2005 Vienna Austria, http://www.itsc2005.at/

• International Journal for Information Technology: International Conference On Software And Knowledge Engineering, (SKE) - 2005: February 25-27, 2005 Istanbul Turkey: http://www.ijit.org/ske2005/organization.htm

• The 7th International Conference on  Enterprise Information Systems will be held at Miami,  May 24-27, 2005 (http://www.iceis.org)

• The 6th International Conference on  Enterprise Information Systems will be held in Universidade Portucalense, Porto - Portugal 14-17, April 2004 (http://www.iceis.org)

• The 7th Maghrebian Conference on Computer Sciences held in Annaba, Algeria, May 6-8th, 2002
  

Session Chair:  

• Smart Engineering System Design Neural Networks, Fuzzy Logic, Evolutionary Programming Complex Systems and Artificial Life (ANNIE 2004), Nov. 7-10, 2004, Marriot Pavilion Hotel, University of Missouri, St. Louis http://web.umr.edu/~annie/annie04/

• The IEEE Midwest Symposium on Circuits and Systems August 14-17 2001, Holiday Inn, Fairborn Ohio.

• Second IEEE Electro/Information Technology Conference held at Oakland University, Rochester Michigan, June 7-9, 2001.

• The 1999 International Conference on Imaging Science, Systems and Technology (CISST'99) held in Las Vegas from June 28th until July 1st, 1999, Associate Editor for the Proceedings of the Conference.

Grants & Awards

Grants & Contracts:

 

• Project ORNOS accepted for funding by the Moroccan Interior Ministry, Starting Jan 2008.

• NSF CAREER: “Adaptive Structural Hidden Markov Models”, recommended for funding by the panel of the Adaptive Artificial Intelligence NSF Program, 2006

• Michigan Space Grant Consortium at University of Michigan at Ann Arbor: Classification of Galaxies via the Galois Lattice Theory, (MSGC-NASA), 2003

• Oakland University Fellowship, 2002

• REU NSF grant, funded in 2002

• French Excellence Grant AUPELF/UREF for a post-doctoral position in University of Quebec at Montreal (1994-95)
  
   

Awards:

 

• Nominated for the Grambling State University Teaching Excellence Award (2009)

• Who's Who in America, Who’s Who in Science and Engineering

• Nominated for the Oakland University Teaching Excellence Award (2006)

• Recipient of the Oakland University Teaching Excellence Award for the Year 2004

• Recipient of the "Naim and Feriel Kheir" Teaching Excellence Award of the School of Computer Science and Engineering for the Year 2004

• Nominated for the Oakland University Teaching Excellence Award (2003)

• Nominated for the SECS (Oakland University) Teaching Excellence Award (2003)  

• Awarded the Oakland University-SECS REF Grant (March 2002)  

• Elevated to IEEE Senior Member (2001)  

• Awarded the AUPELF/UREF French Excellence Grant (1994)

Services 

Editorial Board Member

• Dr. Djamel Bouchaffra is currently a member of the editorial board for the following Journals: 
  
               o    Journal of Pattern Recognition , Editor-in-Chief: Ching Suen,
                     Published  by  Pergamon/Elsevier    http://www.elsevier.com/wps/find/journaleditorialboard.cws_home/328/editorialboard#editorialboard

o   Advances in Artificial Intelligence” (published by Hindawi Corporation) http://www.hindawi.com/journals/aai/editors/

o   The Open Information Systems Journal, Bentham Science Publisher
       http://www.bentham.org/open/toisj/openaccess2.htm

     o   Journal of Engineering Letters Published by the International Association of Engineers.

      http://www.engineeringletters.com/editorial_board_members_p4.html#editorial_members  

o    Journal of Computer Science and Engineering: Scientific Journals International (SJI)

      http://www.scientificjournals .org/editorial_board.htm

• Dr. Bouchaffra has developed the PowerPoint slides of the book titled: "Pattern Classification" by Duda, Hart and Stork, John Wiley & Sons, 2001. His slides are being used by more than 100 Universities throughout the world: visit the following link   
  
  

Guest Editor

Dr. Bouchaffra is a Guest Editor in the following special issues:

•  Special Issue in Advances in Artificial Intelligence Journal (Hindawi), titled: “Machine Learning Paradigms for Modeling Spatial and Temporal Information in Multimedia Data Mining”, Guest Editors: D. Bouchaffra (Lead)  Professor Ce Zhu, Dr. C. S. Chen, Professor  and Dr. A. Amira, March 2011.
  
  

• Special Issue of the Pattern Recognition Journal (Elsevier),  titled: "Feature Generation and Machine Learning for Robust Multimodal Biometrics", Guest Editors: D. Bouchaffra (Lead) and A. Amira, Vol. 41/3, 2008.
  
  

• The Proceedings of the 1999 International Conference on Imaging Science, Systems, and Technology, Editor: H. R. Arabnia, Associate Editors: F. Ahmed, D. Bouchaffra, L. A. Costa, J. T. Favata, R. M. Haralick, Q. Ji, Y. Mun, and M. Sarfraz, Computer Science Research, Education, and Applications Tech. (CSREA) Press, USA, (1999), ISBN: 1-892512-19-X.

Other Services:

• Books Reviewed: “Applied Operating Systems with Java”, by Silberschatz et al., John Wiley & Sons, 2004.

Member of the Peer Review Panel for the Following Governmental Grant Institutions:

• National Aeoronautics and Space Administration (NASA: http://www.nasa.gov/) (ADP Program: Data Analysis and Astrophysics: ROSES-2009 NRA)
  

• Engineering and Physical Science Council, (EPSRC): http://www.epsrc.ac.uk/default.htm United Kingdom (Grants and Proposals)
  

• Indo-US Science and Technology Forum: http://www.indousstf.org/ (Grants and Proposals)
  

Referee for the Following Peer-Reviewed Journals:

• IEEE Journal Transactions on Pattern Analysis and Machine Intelligence (TPAMI)

• IEEE Journal Transactions on Neural Networks (TNN)

• IEEE Journal Transactions on Image Processing (TIP)

• IEEE Transactions on Knowledge and Data Engineering (TKDE)

• Journal of Pattern Recognition published by Elsevier

• Pattern Recognition Letters

• Journal of Knowledge and Information Systems

• Advances in Artificial Intelligence (Hindawi)

He was also a referee in several peer-reviewed conferences in Pattern Recognition such as ICDAR, IWFHR, etc.

IEEE Services and Membership:

• Vice Chair of the IEEE/SEM Computer Section

• IEEE Senior Member

• Member of the IEEE Computer Society

Oakland University Services:

• Chair of the Committee for Academic Standing at Oakland University for the Year 2004

• Member of the Executive Committee of the School of Computer Science and Engineering at Oakland University

• Member of the Undergraduate Curriculum Committee of the Computer Science Department at Oakland University
  

Contributions in Research

At GSU and Oakland University (Michigan, USA):

• The Concept of Topological Hidden Markov Models (THMMs)
  
  Topological features such as distances, shapes have rarely been combined with statistical features. In this research, shape of objects and probability measure are merged together (intrinsically) within a single mathematical framework. Parts of this research has been accepted/published in  different places: "Conformation-based Hidden Markov Models: Application to Human Face Identification", in: IEEE Transactions On Neural Networks (To Appear in 2010); or “Embedding HMM-based Models in a Euclidean Space: The Topological Hidden Markov Models" (oral presentation at IEEE ICPR 2008 conference, Tampa, Florida Dec., 2008). THMMs represent a preliminary attempt to embed Euclidean or even Non-Euclidean spaces (Riemannian geometry) in a HMM graph. For example, by projecting the nodes of a HMM onto a Euclidean space, topological transformations (or mappings) such as homotopies and homeomorphisms will be exploited. Topological HMMs should be able to not only predict a shape of an object but unfold relationships between shapes of objects as well. This topological extension represents a major breakthrough in the capabilities of HMM-based formalisms.
  
  

• The Concept of Structural Hidden Markov Models (SHMMs)
  
  Professor Bouchaffra has introduced the concept of Structural Hidden Markov model (SHMM) which extends the traditional HMMs by allowing structural information to be created and accounted for. A transition probability between structures is defined in order to classify/recognize complex structural patterns such as proteins. Unlike traditional HMMs, SHMMs account for long range dependencies contained in the entire observation sequence (discrete-valued time series). He successfully applied SHMMs in handwritten numeral recognition, where they have outperformed traditional HMMs. He has also applied to mining customer's preferences for automobile designs. The goal is to build a model that aids engineers in their automobile designs. This research that demonstrates the importance of SHMMs in pattern recognition has been published in two premier journals. He has also published 3 peer-reviewed conference papers, one of them was amongst the best paper award. 
  
  

• Functional Set Theory (FST)
  
  Dr. Bouchaffra has also introduced the mathematical concept of Functional Set Theory (FST) that highlights structural patterns. This research which is still in progress has a fundamental aspect since it involves a new axiomatic. It extends fuzzy set theory through the use of a functional relationship between objects in a traditional set. Preliminary results of this novel theory have been published in the proceedings of ANNIE'2005 Conference.   
  
  

• Fusion of Statistics and Logic in Language Modeling
  
  Dr. Bouchaffra devised a new method based on probabilistic logic as a solution to the challenging data (or events) sparseness problem. The goal is to determine a way to assign probability values to testing samples that were not seen in a training set. He applied his approach to language modeling using n-grams as data. He developed two approaches: (i) a topological approach that is based on an invariance property that bounds the set of solutions, and (ii) a Jaines MAXENT criterion for providing with a unique solution. This work has been published in the Pattern Recognition Journal.
  

While he was at CEDAR (Center of Excellence for Document Analysis and Recognition)

State University of New York at Buffalo, USA

• Conversion of Recognition Scores into a-Posteriori Probability values
  

The goal in this research is to combine classifiers with different output scores so that a decision strategy can be computed. This problem is challenging since classifiers output are not in the same scale. Some are distances, others are confidence values, or likelihood numbers. He has designed with an other researcher a new technique which converts recognition scores into a-posteriori probability values. He therefore built a fully Bayesian framework involving signal and language information at a same analysis level. This work has been published in IEEE Transactions PAMI, (see the publication section). 

• NonStationary Markov Chains in ZipCode Recognition
  
  He has introduced and showed the importance of the nonstationarity property of the Markov chains in the recognition of ZIP Codes. This work has been published in IEEE Transactions. PAMI (see the publication section). He has developed a digit recognizer called "PROZE" based on a combination of Fourier descriptors and other spatial features. PROZE (written in C language) is capable to interface with different other programs enabling the user to specify the data structure of the feature vectors during training, the search technique used through the prototype file during testing, etc. The current accuracy of PROZE without thresholding is around 99.2%, when thresholded its accuracy is 99.7%. 

While he was at UQAM (Montreal, Canada)

He has been working on indexing and classification of textual documents. He developed several probabilistic classifiers based on Markov Random Fields. He compared the contribution of connexionist models such as "Adaptive Resonance Theory" (ART) and "Multilayer Perceptron" (MLP) with Gibbs Markov Random Fields. He has published several papers during his postdoctoral position.

While he was at Grenoble University (France)

During his Ph.D. thesis, he has been working on Parts of Speech Tagging for the French language analysis. He has devised several sampling techniques that were part of the Markov chains training phase. He also developed particular hidden Markov models for the French language morphological analysis. His work was published in AAAI conferences and in some mathematical journals.

Students Under His Supervision

Postdoctoral Level:

Name: Franz Pernkopf

Place: Oakland University

Topic:  “Fusion of Genetic Algorithm and Expectation-Maximization (GEM): The GEM Optimization Algorithm” 

(This research was published in IEEE-PAMI Journal and in Conferences)

Ph.D. Theses:  

1.  Name: Nicoleta Rogovschi

    Topic: "Classification à base de modèles de mélanges topologiques des données catégorielles et continues"

    Ph.D. Thesis defended on December 2009

    My Role: Rapporteur" (Reviewer)

    Place: LIPN-UMR 7030 Universit´e Paris 13 – CNRS, Universite’ Paris XIII (France)

2. Name: Jun Tan

    Topic: “Structural Hidden Markov Models and their Applications”

    Ph.D. Thesis, Defended in 2006

    My Role: Principal Supervisor

    Place: Oakland University, Michigan (USA)

3. Name: Krippa Sundar

    Topic: “Design of Language Models for Handwriting Recognition”

    PhD Thesis Defended in 1999

    Place: CEDAR, State University of New York at Buffalo, USA

    My Role: Co-Principal Supervisor

4. Name: Raef Aidibi (3 years)

    Topic: “Structured Hidden Markov Decision Processes”

    My Role: Principal Supervisor

    Place: Oakland University, Michigan

5. Name: Tarek Dakhlallah (3 years)

    Topic: “Improving Hidden Markov Forests”

    My Role: Principal Supervisor

    Place: Oakland University, Michigan

Master Theses and Independent Studies:

1. Name: Dan Zoble

    Place: Oakland University

    Topic: "Implementation of a 3D Fold Protein Recognizer" (Independent Study during fall 2006)

    Place: Oakland University, MI, USA

    Topic: “Intelligent Systems” (Daimler Chrysler Research Project Started in fall 2006)
 

3. Name: Kulkarni, Shruti G.

    Place: Oakland University, MI, USA

    Topic: “Registration of Protein Tertiary Structures” (Master Thesis, Started in winter 2006- Defended in summer 2006)

4. Name: Dariusz Mikulski (online thesis) 

    Place: Oakland University, MI, USA

    Topic: “Classification based on Rough Set Theory” (Master Thesis, Started in fall 2005- Defended in summer 2006)

5. Name: Meha Perashar

    Place:  Oakland University, MI, USA

    Topic: “Generalized Structural Hidden Markov Model” (Independent Study, fall 2005)

6. Name: Shaun Tomasewski

    Place: Oakland University, MI, USA

    Topic: “Recognition of Events in Camera Monitoring Scenes” (Master Thesis, Defended in spring 2005)

7. Name: Praveen Cheruku

    Place: Oakland University, MI, USA

    Topic: “Galaxy Classification using Galois Lattices (The GALILEO System)” NASA funded Project (Independent Study, spring 2004)

8. Name: Arun K. Jinde

    Place: Oakland University, MI, USA

    Topic: “Linguistic Functions of Web-Crawlers: Case study: Google” (Independent Study, summer 2003)

9. Name: Krishna Baisetti

    Place: Oakland University, MI, US

    Topic: “Statistical Functions of Web-Crawlers: Case study: Google” (Independent Study, spring 2003)

10. Name: Kalyani Pendharkar 

     Place: Oakland University

     Topic: “Simulation of an Intelligent Virtual Shopping Cart” (Independent Study, spring 2004)

11. Name: Sawsan Aboul-Hassan

     Place:  Oakland University, MI, USA

     Topic: "Mapping Car Designs to Human Perceptions" (Master Thesis Defended in 2002)

12. Name: Lina Shoshani

     Place: Oakland University, MI, USA

     Topic: "Handwriting Identification" (Master Thesis Defended in 2002)

---

Publications  

New Book: Machine Learning Principles and Adaptive Pattern Classification (To Be Published in 2011)

Journal Papers

The following text applies to any material that appeared in any IEEE Transaction or Elsevier Journal. "Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted components of this work in other works must be obtained from IEEE or Elsevier publishers".

 

1. D. Bouchaffra, "Object Classification via Reconstruction of Alpha-Shapes-based Manifold: Application to Age-Invariant Face Recognition", submitted to: Pattern Recognition Letters, 2011.

   
   

2. D. Bouchaffra, "Mapping Dynamic Bayesian Networks to Alpha-Shapes: Application to Human Faces Identification across Ages", Minor Revision in: IEEE Transactions On Neural Networks and Learming Systems (TNNLS), to appear in 2012.

    

3. D. Bouchaffra, "Conformation-based Hidden Markov Models: Application to Human Face Identification", in: IEEE Transactions On Neural Networks (TNN), vol. 21, no. 4, pp. 597-608, 2010 (Manuscript available in pdf format).
   
   

4. D. Bouchaffra, "Application of Topological Hidden Markov Models to Protein Fold Recognition", Submitted to IEEE/ACM Transactions On Computational Biology and Bioinformatics, 2010 (under revision).
   
   

5. D. Bouchaffra, "Embedding HMM-based Models in a Euclidean Space: The Topological Hidden Markov Models", in: Pattern Recognition, Volume 43, Issue 7, July 2010, Pages 2590-2607 (Manuscript available in pdf format).
   
   

6. D. Bouchaffra and A. Amira, "Structural Hidden Markov Models for Biometrics: Fusion of Face and Fingerprint", in: Special Issue of Pattern Recognition Journal (Elsevier), Vol. 41/3 pp 852-867, 2008. (Among the Top 25 Hottest Articles of the PR Journal) (Manuscript available in pdf format)     
   
   

7. P. Nicholls, A. Amira, D. Bouchaffra and H. Perrot, "A Statistical Multiresolution Approach for Face Recognition using Structural Hidden Markov Models", in: EURASIP: Journal on Advances in Signal Processing: in Special Issue: "Advanced Signal Processing and Pattern Recognition Methods for Biometrics", Publisher Hindawi Corporation, Volume 2008, Article ID 675787, 13 pages.  (Manuscript available in pdf format)
   
   

8. D. Bouchaffra and J. Tan, "Structural Hidden Markov Models using a Stochastic Context-Free Grammar", in: Control and Intelligent Systems, Vol. 35, No. 3, 2007, ACTA Press.
   
   

9. D. Bouchaffra and J. Tan, "Structural Hidden Markov Models using a Relation of Equivalence: Application to Automotive Designs",  in: Data Mining and Knowledge Discovery Journal, Volume 12: 1, Springer-V, 2006. (Springer link) (Manuscript available in pdf format)
   
   

10. D. Bouchaffra and J. Tan, "Structural Hidden Markov Models: Application to Handwritten Numeral Recognition", in:  Intelligent Data Analysis Journal, IDA, Vol., 10:1, IOS Press, 2006. (Manuscript available in pdf format)

11. F. Pernkopf and D. Bouchaffra, "Genetic-based EM Algorithm For Learning Gaussian Mixture Model", in: IEEE Transactions On Pattern Analysis and Machine Intelligence, TPAMI, Vol. 27, no. 8, pp. 1344-1348, Aug. 2005. (Manuscript available in pdf format)

12. D. Bouchaffra, "Probabilistic Logic with Minimum Perplexity: Application to Language Modeling",  in: Pattern Recognition Journal, Publisher: Elsevier, issue 38:8, 2005. (Manuscript available in pdf format) (received a feedback from Professor Fred Jelinek)

13. D. Bouchaffra, "Introduction of Logic in Language Modelling: The Minimum Perplexity Criterion", in: The International Journal of Robotics and Automation, Volume 20, Issue 3, 2005, Publisher ACTA Press.

14. H. Chekireb, M. Tadjine, D. Bouchaffra, "Direct Adaptive Fuzzy Control of Nonlinear System Class: Application to 3-Joint  Robot Manipulator", in: The International Journal of Control and Intelligent Systems, Vol. 31, No. 2, Published by ACTA Press, 2003. (Manuscript available in pdf format)

15. D. Bouchaffra, V. Govindaraju and S. N. Srihari, "Postprocessing of Recognized Strings using Nonstationary Markovian Models",  in: IEEE Transactions On Pattern Analysis and Machine Intelligence, TPAMI Vol. 21, No. 10, October, 1999. (Manuscript available in pdf format)

16. D. Bouchaffra, V. Govindaraju and S. N. Srihari, "A Methodology for Mapping Scores to Probabilities", in: IEEE Transactions On Pattern Analysis and Machine Intelligence, TPAMI Vol. 21, No. 9, Sept 99. (Manuscript available in pdf format)

17. D. Bouchaffra, E. Koontz, V. Kripasundar and R. K. Srihari, "Incorporating Diverse Information Sources in Handwriting Recognition Post-Processing" , in: International Journal of Imaging Systems and Technology (IJIST), vol. 7, issue 4, Published by John Wiley, June 1996, pp. 320-329. (Manuscript available in pdf format) (Link from Wiley)

18. D. Bouchaffra, "Theory and Algorithms for Analysing the Consistent Regions in Probabilistic Logic", in: International Journal of Computers and Mathematics with Applications, Published by Pergamon Press,Vol 25, N.3, pp 13-s25, 1993.  (Manuscript available in pdf format)
    
     

Book Chapters

1. D. Bouchaffra and J. Tan, "Structural Hidden Markov Model and Its Application in Automotive Industry", Enterprise Information Systems V, Camp, O.; Filipe, J.B.; Hammoudi, S.; Piattini, M.G. (Eds.), XIV, 332 p., Hardcover, ISBN: 1-4020-1726-X, Published by Springer, 2004.

2. D. Bouchaffra and J. Rouault, "Capturing Observation in a Nonstationary Hidden Marskov Model" , Selected papers in "Models Selection from Data: Artificial Intelligence and Statistics IV", in: Lecture notes in Statistics, Vol. 89, edit. P. Cheeseman & R. Oldford, published by Springer-Verlag, 1994.

3. D. Bouchaffra, "Consistent Regions in Probabilistic Logic when using Different Norms", Complete Version, in: Selected papers in Artificial Intelligence Frontiers in Statistics: AI and Statistics III, edit. David .J. Hand, published by Chapman and Hall London, pp 370-386, 1993. 
   

4. D. Bouchaffra, "A Relation between Isometries and the Relative Consistency Concept in Probabilistic Logic", Selected papers in "AI, Expert Systems and Symbolic Computing for Scientific Computation", edit. J. Rice and E. N. Houstis, published by Elsevier North-Holland, 1992.

5. D. Bouchaffra and J. Rouault, "A Nonstationary Hidden Markov Model with a Hard Capture of Observations: Application to the Problem of Morphological Ambiguities", in "Probabilistic approaches to Natural Language", Papers from the 1992 Fall Symposium, Technical report FS-92-04, AAAI Press, American Association for Artificial Intelligence, 1993.

 

Papers in Conference Proceedings (Refereed)

1. D. Bouchaffra, "Topological Dynamic Bayesian Networks: Application to Human Face Identification across Ages", (Oral Presentation) in: Proceedings of IEEE Computer Society on Computer Vision and Pattern Recognition Workshops (CVPRW), San Francisco, June 13-18, 2010, pp. 1-8. Manuscript available in pdf format..

2. D. Bouchaffra, "Topological Dynamic Bayesian Networks", in: The Proceedings of The 20th IEEE International Conference on Pattern Recognition (ICPR), August 23-26, Istanbul, Turkey 2010, Manuscript available in pdf format..

3. D. Bouchaffra, "Embedding HMM's-based Models in a Euclidean Space: The Topological Hidden Markov Models", Accepted as oral presentation in: The Proceedings of The 19th IEEE International Conference on Pattern Recognition (ICPR), Tampa Convention Center, Florida, Dec. 8-11, 2008, http://www.icpr2008.org/. Manuscript available in pdf format.

4. P.Nichols, A. Amira, and D. Bouchaffra, "Multiresolution Hybrid Approaches for Automated Face Recognition", Accepted as Oral Presentation in the IEEE NASA/ESA (European Space Agency) AHS Conference, Special Session on Secure Data and Information Systems, August 5-8, 2007, University of Edinburgh, Scotland, UK. Manuscript available in pdf format. http://www.see.ed.ac.uk/ahs2007/AHS.htm.
   

5. Djamel Bouchaffra and Jun Tan, "Protein Fold Recognition using a Structural Hidden Markov Model", in: The Proceedings of The 18th IEEE International Conference on Pattern Recognition (ICPR), Hong Kong, 20-24 August 2006 (Proceedings Published by IEEE Computer Society). Manuscript available in pdf format

6. Djamel Bouchaffra, "Introduction to Functional Set Theory and Its Principles: Fusion of Statistics and Syntax", in: The Proceedings of ANNIE'2004, Smart Engineering System Design-Neural Network, Fuzzy Logic, Evolutionary Programming, Complex Systems and Artificial Life, Nov. 7-10, 2004, University of Missouri-Rolla. Manuscript available in pdf format

7. Djamel Bouchaffra and Jun Tan, "Introduction to Structural HMM and it's Application in Pattern Classification", in: The Proceedings of ANNIE'2004, Smart Engineering System Design-Neural Network, Fuzzy Logic, Evolutionary Programming, Complex Systems and Artificial Life, Nov. 7-10, 2004, University of Missouri-Rolla. Manuscript available in pdf format.

8. Djamel Bouchaffra and Jun Tan, "Introduction to the Concept of Structural HMM: Application to Mining Customers' Preferences for Automotive Designs", in: The Proceedings of The 17th IEEE International Conference on Pattern Recognition (ICPR) Cambridge, United Kingdom, 23-26 August, 2004 (Proceedings Published by IEEE Computer Society). Manuscript available in pdf format.

9. Djamel Bouchaffra and Jun Tan, "Subjective Labeling using a Pyramidal Representation", in: The Proceedings of The 16th International Conference on Industrial & Engineering Applications of Artificial Intelligence and Expert Systems (IEA/AIE-2003), Loughborough, UK, June 23-26, 2003.

10. Djamel Bouchaffra and Jun Tan, "Structural HMM Modeling and its Applications in Automotive Industry", in: The 5th International Conference on Enterprise Information Systems, Angers - France 23-26, April 2003.

11. Djamel Bouchaffra, and Jun Tan,s "Mapping Designs to User Perceptions using a Structural HMM: Application to Kansei-Engineering", in: IEEE International Conference on Computational Intelligence for Modelling, Control and Automation - CIMCA'2003 (chaired by Lotfi Zadeh and Stephen Grossberg), 12-14 February 2003, Vienna - Austria.

12. Sawsan Aboul-Hassan and Djamel Bouchaffra, "Automotive Design Driven by Pattern Recognition", in: Smart Engineering System Design, Annie'2001 Conference Proceedings, University of Missouri-Rolla, November 4-7, 2001 (nominated for the best paper award).

13. M. Zohdy, D. Bouchaffra and J. Quinlan, "Optimal Mapping from Chromosome Space to Feature Space for Solving Sequential Pattern Recognition Problems", in: Proceedings of the IEEE Midwest Symposium on Circuit and Systems, Dayton, Ohio, August 14-17th, 2001.

14. Lina Shoshani and Djamel Bouchaffra, "Identification of Handwritten Digits Using K-Nearest Neighbor", in: Proceedings of the Second IEEE Electro-Information Technology Conference, Oakland University June 7-9, 2001.

15. D. Bouchaffra, V. Govindaraju and S. N. Srihari, "Recognition of Strings using Non-stationary Markovian Models: An Application to ZIP Code Recognition" , (Paper Accepted as Oral Presentation) in: Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'99), June 23-25, 1999, Fort Collins, Colorado, USA.

16. D. Bouchaffra, V. Govindaraju and S. N. Srihari,"A Methodology for Deriving Probabilistic Correctness from Recognizers", (Paper Accepted as Oral Presentation) (on-line pdf format) in: Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'98), Santa Barbara, California, USA, June 23-25, 1998.

17. Djamel Bouchaffra, Eugene Koontz, V. Kripasundar, Rohini K. Srihari and Sargur N. Srihari, "Integrating Signal and Language Context to Improve Handwritten Phrase Recognition: Alternative Approaches", (on-line ps format) in: Proceedings of the Sixth International Workshop on Artificial Intelligence and Statistics, Ft. Lauderdale Florida USA, Jan. 4-7, 1997.

18. D. Bouchaffra and J.G. Meunier, "A Markov Mesh on Uncertain-Galois Lattice: Classification in Terminology", (on-line pdf format) in: Proceedings of the "3rd International Conference on Statistical Analysis of Textual Data, SATD'95, Rome, Italy, December 11-13., Published by CISU.  

19. D. Bouchaffra and J.G. Meunier, "A Thematic Knowledge Extraction Modeling through a Markovian Random Field Approach" (on-line pdf format) in: Proceedings of the 6th International DEXA Conference and Workshop on Database and Expert Systems Applications, DEXA'95, Sept 19-22, London, U.K., Published by DEXA Association.  

20. D. Bouchaffra D. and J. G. Meunier, "A Markov Random Field Approach to Automatic Information Retrieval", (on-line pdf format) in: Proceedings of the "Third International Conference on Document Analysis and Recognition, ICDAR'95", August 14-16, Montréal, CA, published by IEEE Society (nominated for the best paper award).

21. D. Bouchaffra, G. Lallich-Boidin and J. Rouault, "Champs de Markov et Espaces Métriques dans la Recherche d'Information",  in: Proceedings of the ALLCACH, 18-22 avril 1994, Sorbonne, Paris (France).

22. D. Bouchaffra, G. Lallich-Boidin and J. Rouault, "Echantillonage Stratifié avec Restimation (Bootstrap): Application à l'Analyse Morphologique", in: Proceedings of the Secondes Journées internationales d'Analyse Statistique de Données Textuelles, Montpellier (France), 21-22 octobre 1993.

23. D. Bouchaffra and J. Rouault, "Different Ways of Capturing the Observations in a Nonstationary Hidden Markov Model: Application to part of speech tagging", in: Proceedings of the Fourth International Workshop on Artificial Intelligence and Statistics, Fort-Lauderdale, Florida USA, Jan. 4-7, 1993.  

24. D. Bouchaffra, "Consistent Regions in Probabilistic Logic when using Different Norms", in: Proceedings of the Third International Workshop on Artificial Intelligence and Statistics, Fort Lauderdale, Florida, USA, Jan. 2-5, 1991.   s

25. D. Bouchaffra, "A Relation between Isometries and the Relative Consistency Concept in Probabilistic Logic", in: 13th IMACS world Congress on Computation and Applied Mathematics, Trinity college, Dublin, Ireland, july 22-26, 1991.
    

---

Research Papers In Preparation

1. D. Bouchaffra, "Embedding Dynamic Bayesian Networks in Euclidean and Non Euclidean Spaces", 2011

2. D. Bouchaffra, "Topological Markov Decision Processes", 2011.

---

Teaching Publication

1. D. Bouchaffra, "Teaching: An Art But A Mutual Challenge", Published in the Oakland University Newsletter,  March 2005.

---

Dr. Bouchaffra's Cited Articles

 

 For a complete list of citations of his research, visit http://scholar.google.com/scholar?start=0&q=author:Djamel+author:Bouchaffra&hl=en&as_sdt=0,5

---

External Comments on Dr. Bouchaffra's Research and Teaching  

Dr. R. Kasturi (President Elect of the IEEE Computer Society, former Editor-in-Chief: IEEE TPAMI): "I have known Dr. Djamel Bouchaffra through his publications in journals since 1993. He has made important contributions in the research area of pattern recognition, in general, and handwriting recognition, in particular. Dr. Bouchaffra has published several papers in journals and peer-reviewed conference proceedings. In particular, his papers have been published in the premier journal of our field, IEEE Transactions on Pattern Analysis and Machine Intelligence and in the premier conference in our filed, IEEE Computer Vision and Pattern Recognition (CVPR). I was the session chair at one of his CVPR presentations. His presentation was well received by the audience. This work is a significant contribution that advanced the handwritten digit recognition research for the specific domain of postal zip code recognition. I am pleased to see that he has papers submitted to PAMI and CVPR this year. I am also pleased to note his interest in the fusion of statistics and syntax in pattern recognition. This approach, if widely accepted, would have many applications. I am unable to comment on the significance of his other recent contributions since they are outside the scope of my primary research interest."

Dr. A.  Zemirline (Associate Professor at Brest University): "Doctor Bouchaffra is a very active researcher who has brought a great deal of enthusiasm, experience and know-how to the state-of-the-art. His knowledge of pattern recognition and image vision algorithms and techniques is outstanding. Prof. Bouchaffra is an expert in statistical and mathematical modeling as well as, learning machines, artificial intelligence and soft computing. His has developed mathematical models that are capable to: (i) represent optimally structural patterns, (ii) decide under uncertainty, and (iii) classify these complex patterns. His areas of application involve speech/handwriting recognition, language modelling, data mining, and bioinformatics. Prof. Bouchaffra has introduced the concept of (SHMM) which extends the traditional HMM's by allowing structural information to be created and accounted for. He successfully applied SHMM's in handwritten numeral recognition, where they have outperformed traditional HMM's. He has also applied to mining customer's preferences for automobile designs. The goal is to build a model that aids engineers in their automobile designs. Two papers that show the role of SHMM's in pattern recognition have just been accepted for publication in: The Journal of Pattern Recognition and in: The International Journal of Data Mining and Knowledge Discovery. These two journals are considered premier in their fields. He has also developed an efficient algorithm that combines expectation-maximization with genetic algorithms. This paper has been published in IEEE PAMI, which is also premier journal in the field. He has published 3 peer-reviewed conference papers; one of them was amongst the best paper award. Prof. Bouchaffra has also introduced the concept of functional set theory (FST) in which preliminary results have been published in the proceedings of ANNIE'2005 Conference. He is also working on statistical language modeling where he has merged statistics with logic to solve the n-grams sparseness problem. This work has been published in the Journal of Pattern Recognition."

Dr. D. Hanna (Assistant Professor at Oakland University, Michigan): "I have had a unique opportunity to be both a student in one of Dr. Bouchaffra's courses (a graduate course in applied pattern recognition) and a colleague. I received my Ph.D. from Oakland University in 2003 and successively joined the faculty. Dr. Bouchaffra is an excellent teacher and is always willing to provide help to students who need it and go further in depth on topics fro students who are interested in learning more. Dr. Bouchaffra and I have also had several research-related discussions over the past three years about new pattern recognition techniques and their potential applications to biological problems and implementation in embedded systems. Dr. Bouchaffra is not only well-versed in computer science, but he also has an extremely solid theoretical background which enables him to quickly take problems to very abstract, fundamental levels and consider new theories from which solutions could be derived. Dr. Bouchaffra has a high capacity and talent for forming and solving theoretical conjectures using complex mathematics for applications in computer science. It has been a pleasure to take Dr. Bouchaffra's course and work with Dr. Bouchaffra."

---

A Sample of Developed Systems

Currently he is working on analyzing the effect of growth and aging on the facial topology for identification 

"FACE-MORPHOTRANS": A program based on topological and structural HMM's that is capable to extract facial features that are more affected by growth and aging for prediction as well as invariant features that are less sensitive to aging for human identification.

 

 

While at Oakland University: He has designed and implemented the following systems:

                

	“BIOSTRUCT”: A face recognition system that uses a structural hidden Markov models (SHMMs). It has been implemented in MATLAB and tested on the FERET and AT&T databases.
	“HELPCOP”: A program that is capable to identify frames of a video sequence that contain highway patrol troopers in difficult situations.
	"PROTCLASS": Software that performs a classification of amino-acid sequences into 3D protein folds.
	"GALILEO": A graphical user-interface captures an image of a galaxy from the Sloan Digital Sky Survey (SDSS), extracts features from this image and classifies the galaxy using Galois Lattice theory. It also performs a browsing of the sky, extracts learning rules and provides an inventory of the galactic system. For more information, visit: http://www.redplanet5.net/galileo
	"PREDICAR": Software that helps design engineers to predict customer's preferences in automotive industry.
	"NEXTWEB": A web-crawler that embeds the lexical database WordNet in order to capture semantic relationships between words and use statistical properties. This system represents a fusion between linguistics and statistics.

While he was at CEDAR (State University of New York @ Buffalo)

	PROZE  (PROfiles, Zoning using Exterior Contour) A Recognizer of Single Digit Images Click here to view the PROZE GUI
	ZPP (Zipcode Post-Processor)   A Recognizer of US ZIPCODES (sequences of 5 digit images)
	FNM (Firm Name Matcher)   A Firmname Recognizer: It matches a handwritten company name in a mailpiece with a list of company names stored in a large size lexicon.

 

The recognition rate of PROZE for single digit is around 99%. The recognition of US ZIPCODES is based on a nonstationary HMM and uses contextual information between digits. Besides, the two "CheckButtons" that you see in the GUI image enables the user to switch between the two modes "Optimized" and "Nonoptimized" of PROZE. The mode "Optimized" clusters the prototype using EM and K-means clustering schemes and increase the speed of the digit or ZIPCODE recognition time per image. The ZIPCODE recognition work has been published in IEEE Transactions PAMI (see section Journal Papers)

 

---

Invited Talks/Seminars

 

Dr. Djamel Bouchaffra has been invited internationally to present and discuss his research. Papers or Powerpoint slides of the following talks can be accessed on demand by sending email to: dbouchaffra@ieee.org

• "Structural and Topological Hidden Markov Models: Dynamic Bayesian Network Embedding"
  Place: Institut Galile'e (apprentissage non supervise a base de mode`les de Markov cache's structure's et topologiques) 
  (LIPN UMR 7030 du CNRS, Universite Paris XIII, Nord, France), Prof. Invited in June 2010.

• "Concept Formation based on Galois Lattices: Application to Galaxy Classification"
  Place: University of Michigan at Ann Arbor, Oct. 11th, 2003, invited by the Michigan Space Grant Consortium (NASA sponsor).

• "Applied Research Review in Pattern Recognition",
  Place: Oakland University, MI, October 10th, 2003 invited by Michigan Industrial Partners

• "Mining Customers' Preferences for Automotive Designs",
  Place:  Southfield, Michigan, March 26th 2003, invited by the IEEE Chapter V.

• "Solving the Sparse Events Problem through Probabilistic Logic:  Application to Speech Recognition"
  Place: Department of Mathematics, Oakland University, USA, March 6th 2003, invited by the Department Chair.

• "Combining Algebraic/Topological Structures with Stochasticity: Application to Classification"
  Place: School of Computer Science and Engineering, Oakland University, USA, March 20th 2002, invited by Prof. C. Kobus.

• "Mapping Objects Designs to Human Perceptions"
  Place: Electrical and Systems Engineering Department, Oakland University, USA, Nov. 7th 2001, invited by Prof. Kheir (chair of the ESE).

• "Pattern Recognition In Practice"
  Place: Delphi Automotive, (Troy) Michigan State, USA, Feb. 12th 2001, invited by Dr. B. Maiteh
  Place: School of Computer Science and Engineering, Oakland University, USA, Nov. 17th 2000, invited by Prof. Loh.

• "Incorporating Signal and Language Sources in a same Bayesian Framework"
  Place: CS department, State University of New York at Buffalo, Amherst, NY, USA, Feb. 6th 1998, invited by Prof. S. Shapiro.

• "A Dynamical Information Retrieval System using A Gibbs-MRF distribution", Congres ACFAS'95
  Place: Chicoutimi, Quebec, Canada, 1995.

• "Hidden Markov Models and Knowledge Extraction", Universite du Quebec A Montreal
  Place: Montreal, Quebec, Canada, June 1994.

• "The Contribution of Hidden Markov Models in Natural Language Processing", Universite du Quebec A Trois Rivieres
  Place: Trois-Rivieres, Quebec, Canada, March 1994, invited by Prof. Sylvain Delisle at the CS department.

• "A Sampling Procedure and a Vectorial Markov Chain for Parts of Speech Tagging"
  Place: Rank Xerox Research Center, Grenoble, France, Dec. 6th 1993, invited by Dr. Jean Pierre Chanod.

• "Les Chaines de Markov Dans CRISTAL: Seminar on Applications of Markovian Stochastic Processes"
  Place: Institut National de Telecommunication, INT, Ivry Paris, France, June, 15th 1993.

• "Probabilistic Logic and Meta-Knowledge",
  Place: Universite Joseph Fourier: La Tronche, Grenoble, France, Jan 1992, invited by Prof. Vincent Rialle. 

  ---

Other Related Information Servers

• American Mathematical Society

• International Association for Pattern Recognition

• Association for Uncertainty in AI

• ASA Section on Bayesian Statistical Sciences

• International Society for Bayesian Analysis

• Knowledge Discovery Mine .

• Machine Learning Information Services

• Minimum Message Length Encoding

• Pattern Recognition Information

• Software for Belief Networks

• SIGART Electronic Information Service

• Statistics servers "maintained at" Institute of Statistics and Decision Sciences, Duke University

• Statlib Index "a system for distributing statistical software, datasets, and information maintained at CMU"

• Bayesian's Worldwide

• Mixture Modeling "maintained by David Dowe.

---