×
Riadh Fezzani

Riadh Fezzani

Doctor in applied mathematics
154 avenue Marx Dormoy, 92120, Montrouge. France
+33 6 75 86 21 97
rfezzani@gmail.com
https://rfezzani.github.io
French Arabic English

Background

About

About

Autonomous and curious, 8 years of experience including a Ph.D., I appreciate multidisciplinary environments and the implementation of efficient algorithms.
I support open source projects and promote the KISS principle.

Skills

Skills

  • Programming
    C/C++ Python Matlab HTML5/javascript Qt Git
  • High performance computing
    Multithreading GPGPU openMP CUDA openCL Metis/Scotch
  • Image processing
    Optical flow Registration Stitching Template matching Segmentation Active contours openCV Matplotlib scikit-image
  • Linear algebra
    Dense and sparse systems Direct and iterative solvers Preconditionning Low-rank approximation BLAS/LAPACK Numpy/Scipy
  • 3D
    Ray tracing Shading openGL/GLSL Blender VTK
Work Experience

Work Experience

  • Engineer,  Ausy

    Since November 2014

    I actually work in the R&D departement of Océ - A Canon compagny, within the Eiger project (high resolution, large format and full color 2.5D and 3D printing). My work consists in developping multiple tools for the visualization and the manipulation of digital 3D objects. During this mission, I participated to three patent filings.
    • Mesh voxelization
    • SVO data structure
    • Volume slicing
    • Realtime visualization
    • Photorealism
    • Height map and texture generation
    • Cutpath generation
    • Blender plugin

  • Postdoctoral research,  IOGS

    June 2013 - November 2014 18 monthes

    Within the SPIM team in the laboratoire Charles Fabry, I worked in the context of the Reveal project in close collaboration with the iMove team at ONERA, the Institut de la Vision and Imagine Eyes on the design of efficient algorithms for the processing of eye fundus images aquired with an adaptive optics camera.
    • Code acceleration and optimization
    • Fast image registration
    • Image enhancement
    • Acquisition area detection (mutual information maximization)
    • Blood vessel activity detection
    • UI for data base manipulation

  • Postdoctoral research,  INRIA

    March 2011 - May 2013 2 years

    Within the PARALL team in the LRI, I worked in the context of the PetalH project on the design of new parallel preconditioners made for the acceleration of the resolution of large ill-conditioned sparse linear systems.
    • Sparse linear algebra
    • Iterative linear solver
    • Preconditionning
    • Sparse matrix graph reordering
    • Domain decomposition
    • Approximate block LDU factorization with filtering constraints
    • Low-rank approximation

  • Ph.D. thesis,  ONERA

    November 2007 - January 2011 3 years

    In the hight performance computing team (DTIM/CHP), “Parallel approach for optical flow estimation by variational method” (Advisors: Frédéric Champagnat and François-Xavier Roux).
    • Combined local gloal approach
    • Total variation regularization
    • Picard/SOR solver
    • Color labeling
    • Variable splitting
    • High performance computing (multithreading, GPGPU)
    • Particle image velocimetry

  • Internship,  CERFACS

    April 2007 - September 2007 6 monthes

    In the parallel algorithms team, in collaboration with ICL from the university of Tennessee: “Study of the parallelization of the \(LDL^T\) factorization in the symmetric indefinite case” (Advisor: Marc Baboulin).
    • Dense Linear Algebra
    • Direct linear solver
    • \(LDL^T\) factorization
Education

Education

  • Ph.D.,  Université Pierre et Marie Curie (Paris VI)

    November 2007 - January 2011

    Applied mathematics

  • M.D.,  Université Paul Sabatier (Toulouse III)

    September 2006 - July 2007

    Mathematical modeling and computer science
Volunteer Work

Volunteer Work

  • Journal and conference reviewer, OSA, IPDPS, PARCO

    Actually

  • Rugby coach, COMB Rugby

    Since September 2010

Publications and communications

Publications and communications