Karim T. Abou-Moustafa

The Artificial Perception Laboratory
McGill Centre for Intelligent Machines
Dept. of Electrical and Computer Engineering
McGill University

C.V. |  Vita |  Research |  Publications |  Graduate Studies |  Awards |  Contact

[Vita]

I am a Ph.D. candidate in the department of Electrical and Computer Engineering and the Centre for Intelligent Machines (CIM) at McGill University under the supervision of Frank Ferrie "Elder Denizen" (director) of APL.

Prior to that, I was a research assistant at The Institute for Research in Immunology and Cancer (IRIC) and the Lab. d'Informatique des Systèmes Adaptifs (LISA) at Université de Montréal. I finished my Masters degree in Computer Science at Concordia University, Montréal, QC, where I was a member of the Centre for Pattern Recognition and Machine Intelligence (CENPARMI). For my undergraduate studies, I obtained my B.Sc. in Computer Engineering from The Arab Academy for Science and Technology (AAST), Alexandria, Egypt.

News

As of April 3rd 2009, I am in an extended research visit to the Component Analysis Lab. at The Robotics Institute of Carnegie Mellon University, Pittsburgh, PA.

I submitted my Ph.D. thesis proposal on March 31st 2009.

[Research]

The main theme of my research is statistical inference and statistical analysis of data. This is the main reason behind my growing interest in the machinery of statistical learning algorithms. Usually, my research leads me further afield, and I end up exploring other interesting areas such as statistics, probability, numerical optimization, numerical linear algebra, and information theory (very recently). On the application frontier, I had the chance to grasp a considerable experience in handwriting recognition and peptide modeling. I have also developed a comprehensive knowledge in fundamentals of computer and biological vision, and video processing (recently). For more information, you can check my Graduate Studies section.

Research Interests

Machine learning algorithms and statistics

Hidden Markov models (HMMs)
Combined Generative-Discriminative models
Time series (or sequential) data classification and clustering
Supervised learning algorithms for classification
Metric learning (supervised and semi-supervised)
Manifold learning algorithms
Local learning and kernel based algorithms
Unsupervised learning (clustering and nonlinear dimensionality reduction)
Density estimation (parametric and non-parametric (kernel-based))

Applications

Offline and online handwriting recognition
Proteomics (peptide modeling and detection)
Computer vision (object recognition)
Video processing (video surveillance mining, temporal segmentation, event and action categorization in video)

[Publications]

Refereed Publications

• K. Abou-Moustafa and F. Ferrie, "Regularized Minimum Volume Ellipsoid Metric for Query-based Learning",
  To appear in the 7th International Conference on Machine Learning and Applications, San-Diego, California, 2008.


• K. Abou-Moustafa and F. Ferrie, "Local Metric Learning on Manifolds with Applications to Query-based Operations",
  To appear in the 7th International Workshop on Statistical Pattern Recognition, Orlando, Florida, 2008.


• K. Abou-Moustafa and F. Ferrie, "Fast and Regularized Local Metric for Query-based Operations",
  To appear in the 19th International Conference on Pattern Recognition, Tampa, Florida, 2008.


• K. Abou-Moustafa and F. Ferrie, "The Minimum Volume Ellipsoid Metric",
  Lecture Notes in Computer Science (LNCS) 4713, Pattern Recognition - 29th DAGM Symposium,
  F. Hamprecht, C. Schnorr and B. Jahne (Eds.), pp. 335 - 344, Springer, 2007.


• K. Abou-Moustafa, M. Cheriet, and C. Suen, "Classification of Time-Series Data Using a Generative / Discriminative Hybrid",
  IEEE Proc. of the 9th International Workshop on Frontiers in Handwriting Recognition (IWFHR), pp. 51 - 56, 2004.


• K. Abou-Moustafa, M. Cheriet, and C. Suen, "A Generative-Discriminative Hybrid for Sequential Data Classification",
  Proc. of the IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Vol. 5, pp. 805 - 808, 2004.


• K. Abou-Moustafa, M. Cheriet, and C. Suen, "On The Structure of Hidden Markov Models",
  Pattern Recognition Letters, Vol. 25, pp. 923 - 931, June 2004.

Theses

• Karim Abou-Moustafa, "A Generative-Discriminative Framework for Time-Series Data Classification",
  Masters Thesis, Concordia University, Montréal, QC, Canada, 2004.


• Karim Abou-Moustafa, "Offline Recognition of Handwritten Arabic and Hindi Digits",
  Bachelor Thesis, Arab Academy for Science & Technology and Maritime Transport, Alexandria, Egypt, 1999.

[Graduate Studies]

My Ph.D. Committee is formed of three members; Prof. F. Ferrie (artificial perception and computer vision), Prof. J. Clark (early vision and attention models), and Prof. G. Dudek (robotics). I submitted my Ph.D. thesis proposal in March 2009, and I passed my comprehensive exam in Aug. 2007. My reading list covered the following topics:

• Fundamentals of computer vision and its statistical approaches  (F. Ferrie) Reading List
  How all computer vision problems are ill posed inverse problems


• Early vision and attention models  (J. Clark) Reading List
  An eye opener reading on how vision is a hard problem


• Statistical models for simultaneous localization and mapping (SLAM)  (G. Dudek) Reading List
  Basically, it covered Kalman filters and Sequential Monte Carlo for SLAM

During the course of my Master's and Ph.D. studies I took the following courses:

• Design and analysis of algorithms (Concordia, CS. Dept.)


• Pattern classification I (Concordia, CS. Dept.)


• Pattern classification II (Concordia, CS. Dept.)


• Neural networks (Concordia, ECE. Dept.)


• Fundamentals of machine learning (Univ. of Montréal, DIRO)


• Optimization and optimal control (McGill, ECE Dept.)


• Mathematical foundation of systems (McGill, ECE Dept.)


• Introduction to computer vision (McGill, SoCS)

I also had the opportunity to audit the following courses at McGill University:

• Mathematical statistics I (A. Vandal, Dept. of Mathematics and Statistics)


• Matrix computations (X-W. Chang, School of Computer Science)


• Numerical estimation (X-W. Chang, School of Computer Science)


• Multi-Disciplinary optimization (S. Nadaraja, Mechanical Eng. Dept.)

[Awards]

• FQRNT - REPARTI Award for International Training

  From "Le Fonds Quebecois de la Recherche sur la Nature et les Technologies" (FQRNT), and "Le Regroupement strategique pour l'etude des Environnments Partages Intelligents" (REPARTI).



• PRECARN Student Scholarship

  From The Institute of Robotics and Intelligent Systems (IRIS) and PRECARN Incorporated.

[Contact]

3480 University street,
McConnell Engineering Building, Room 410,
McGill Centre for Intelligent Machines,
Montréal, QC, H3A 2A7, Canada

Office : McConnell Engineering Building, Room 333

Phone : +1 (514) 398-1282     Fax : +1 (514) 398-7348