Abdulrahman Oladipupo Ibraheem

I studied  computer science and engineering at Obafemi Awolowo University (OAU), Ile-Ife, where I am a member of the Computing and Intelligent Systems Research Group. Before coming to OAU, I obtained secondary school education at Igbobi college, Lagos. There, I finished as the third best student in the 2000 graduating class!

My interests span human and computer vision, machine learning, mathematics and algorithms. I seek techniques, algorithms, models and, ultimately, theories that facilitate the realization of smarter intelligent sytems, especially as far as the visual modality is concerned. My work in the computer vision/ machine learning domain began in 2011 with my undergraduate project. In that project, using an hybrid of a Bayesian classifier and a decision tree,  I sucessfully developed a character recognition system for the diacritized uppercase Yoruba letters.  We presented this work at the AfLat 2011 conference and a  draft of a paper on it can be found under the publications section below.
In a most recent research endeavour, I have proposed a novel feature extraction technique which I call "SENNS" (Sparse Extraction Neural NetworkS), and pronounce "SENSE." The proposed technique is based on the minimisation of an objective function which I formulated. To minimise the function, I have sucessfully derived a gradient descent algorithm which employs both the Standard Backpropagation Algorithm (SBA) as well as a variant of the SBA which I derived myself via rigorous pencil-and-paper mathematics. The details of my workings can be found at:   http://arxiv.org/pdf/1412.6749v1.pdf
On the non-mathematical side, I have developed a prototype plate number recognition system; in retrospect, I see that the inner-workings of my system can be viewed as constrained optimisation framework. To the best of my knowledge, my system is the first ever of its kind to be built on Nigerian soil. Please click on the link  "Plate Number Recognizer" near the upper left of this page to download the system, and instructions for testing it in the MATLAB environment.
The "Docums" below are not publications; they are some of my writing samples, which reflect the work I have been doing ( See the bottom of this page for my publications):

  • Docum1: Curriculum Vitae (CV) pdf


  • Docum2: Ibraheem A.O. (2014) Minimization of Functionals of Diffeomorphisms for 2-d Shape Matching pdf


  • Docum4: Ibraheem A.O., Okusanya A.O. and Odejobi O.A. (2012) Recognition of Handwritten Lowercase Yoruba Letters using  Nearest Neighbour Classifier and Orthogonal Fourier-Mellin Moments.  pdf of abstract


  • Docum5: Ibraheem A.O. (2012). A Concept Paper on a New  Model for Automatic Diacritic Restoration. pdf


  • Docum6:  A Derivation of a Formula in the Tutorial Paper of Lawrence Rabiner on Hidden Markov Models. pdf


  • Docum7Illustration of Learning in Artificial Neural Networks usiing the Perceptron as a Case-Study. pdf


  • Docum8A Recurrence Formula Solves a Theoretical Problem about  the Axis of Least Inertia of a Two-Dimensional Object. pdf



  • Ibraheem A.O. (2014) ALIASE: Application of Axis of Least Inertia for Skeleton  Extraction in Handwritten Character Images.  Accepted with Revisions to Elsevier's Journal of Visual Communication and Image Representation . pdf of draft


  • Ibraheem A. O. (2014). “SENNS: Sparse Extraction Neural NetworkS for Feature Extraction” In Preparation for Springer’s Journal of Mathematical Imaging. pdf of draft


  • Ibraheem A. O. (2014).  “Correlation of Data Reconstruction Error and Shrinkages in Pair-wise Distances under Principal Component Analysis. ” In Preparation for Journal of  the Royal Statistical Society. pdf of draft


  • Ibraheem A.O. and Odejobi O.A. (2011) "Exploiting Handwritten Character Recognition Systems in African Languages Documentation and Education. Oral Presentation at AfLaT 2011. " pdf"