anisbenyelloul'shomepage

About me, my interest, my research...

Contact: anis [dot] benyelloul [at] imag [dot] fr
Curriculum Vitae: (PDF)

I am currently a Ph.d Student at the Joseph Fourier University in Grenoble, France.

I've been introduced to computers and programming by my parents (who have a computer science education too) when I was at high school, and have never lost interest since then.

Over the years, I kept studying computers (in my spare time) and especially computer programming, computer languages and later software architecture. In the meanwhile, I discovered GNU/Linux and the existing world of free software and became addicted to the idea of sharing software just like researchers share their work with the world for the benefit of mankind.

Research Interests

Service Oriented Architecture
Software Engineering Methods (Agile software development, Unified Process, Extreme Programming, etc)
Computer Languages (C++ mainly)

Past Works and Publications

Protocol Oriented Service Deployment Platform

This project has been conducted during school year 2008/2009 and was a partnership between Joseph Fourier University (Grenoble, France) and the National Institute of Informatics (Tokyo, Japan). I have presented this project as my master thesis:

Author: Anis Benyelloul (Joseph Fourier University, Grenoble France) anis.benyelloul [at] gmail [dot] com.
Supervisor: Dr Fuyuki Ishikawa (National Institute of Informatics, Tokyo, Japan) f-ishikawa [at]nii.ac.jp

Abstract

Service Oriented Computing is gaining momentum and interest both in the industry and in the research domain as the paradigm for the next generation of business software architecture. In this context, one particular aspect has been receiving much attention lately: service protocols. Service protocols are enhanced interface descriptions that contain not only the enumeration of the possible messages that a service can send and receive but also the sequencing constrains that must apply for a conversation to be valid. The purpose of this work is to show how the idea of service protocols can be leveraged at runtime, and how to design a ``protocol oriented service deployment platform'' that implements those concepts.

Protocol Oriented Service Platform Proposal (PDF | LaTeX) is the initial proposal that was the starting point of the project.
SOA, Beyond the Hype (PDF | LaTeX) an introductory paper about service oriented architecture. Here I discuss what consitutes the real innovation behind the idea of service oriented computing, and how does it differ from previous paradigms.
Protocol Oriented Service Platform Thesis (PDF | LaTeX) is the thesis report.
Master Thesis Presentation (SWF) This is the slides I used for the presentation. In order to ``play'' them, just click anywhere in the middle of the slides. The arrows at the bottom are for ``fast forward'' and ``rewind''.

Image Segmentation Using Biomimetic Algorithms

"Image Segmentation using Biomimetic Methods" is an end term project presented at the l'INI (l'Institut National d'Informatique, Algeria) for the purpose of obtaining first graduation as engineer in computer science, school year 2006/2007. The project has been realized by:

Anis Benyelloul (National Institute of Computing, Algeria) anis.benyelloul [at] gmail [dot] com
Said-eddine Bensalem (National Institute of Computing, Algeria) said.bensalem [at] gmail.com

The project has been proposed and supervised by:

Dr. Koudil (National Institute of Computing, Algeria) (m_koudil [at] ini [dot] dz)
Dr. Karima Benatchba (Institut National d'Informatique) (k_benatchba [at] ini [dot] dz)
Dr. Brahim Souici (Institut National d'Informatique) (b_suici [at] ini [dot] dz)

Abstract

The work described in this report, is the study of methods inspired from natural behaviors for the image segmentation problem. For this study, we have chosen two biomimetic algorithms: Ainet and PSO. Ainet (stands for Artificial immune network), is an algorithm inspired from immune reactions of vertebrates, while the PSO (which stands for Partical Swarm Optimisation) is a population-based algorithm inspired by social behavior of bird flocking.

In this report, image segmentation is considered as a data clustering problem. Thus, the studied algorithms are modeled to fit with a problem of pixel classification.

The Ainet algorithm uses a learning-based method to separate pixel classes in the image. While PSO considers the segmentation as a combinatorial optimization problem, with a goal of finding an optimal solution, which represents a set of class centers minimizing a fitness function.

The methods have been tested on medical images, and compared with two well-known clustering algorithms: K-means and fuzzy-C-means. Moreover, the Ainet and PSO have been hybridized with fuzzy-C-means in order to enhance the quality of segmentation results.

Keywords :Segmentation, data clustering, optimization, biomimetic algorithms.

uHerbert: Robot Programming Game

uHerbert home page

A few years ago I took part in an international student contest (ImagineCup) which involved (among others) an algorithm challenge. The algorithm challenge consisted of playing a ``robot programming game'' named Herbert, where you have to write a programme in a specific language (called h) to guide the movements of a robot on a level. The contest included a downloadable version of the game that needs to contact the contest servers to download the levels. At that time I didn't have a full-time access to an internet connection so I couldn't use the provided game. So I took it upon myself to write a clone of the official game, that would implement the same rules and look the same, but that would be able to work in offline mode. uHerbert was the result: a free software, offline version of the official ImagineCup Herbert implementation. uHerbert managed to grab some attention when I posted about it on the ImagineCup forums, and I hope it had helped other people as well. The uHerbert home page contains the original program sources and binaries as well as a tutorial on the game itself and the programming language used in it.