Personal tools

Mohammed S. Al-Rawi


Jump to: navigation, search

Status Former member
Country Portugal
Former PhD students
Former masters students
Past projects
Events CAIP2011, ECCB12, ENS 2011, HBM 2011, ICIAR 2012, ICIP 2012, ISI proceedings
Proposals GOLD - Genome Wide Association, PermuTest
Personal page



My current research is a mix between pattern recognition and neuroimaging.The purpose is using machine learning and pattern recognition algorithms on functional magnetic resonance imaging data to decode brain states, and/or analyzing structural brain data. I am also interested in developing computer vision techniques to detect 2D and 3D objects, applying evolutionary algorithms, and developing/using permutation testing methods. Besides music, (soccer) video games, scientific research is my other favorite pastime.


- Functional and Structural (human) Brain Analysis 
- Medical Image Processing
- Multi-voxel Pattern Analysis
- 2D and 3D object recognition, invariant to affine and rigid transformations
- Pseudo Zernike and Zernike Moments
- Multi-objective Optimization using Evolutionary Algorithms / Genetic algorithms
- Higher Order Neural Networks
- Pattern Classification
- Machine Learning
- Diabetic Retinopathy
- Genome wide association studies
- Music Generation via Computer Algorithms
- Iris Recognition and Clustering
- Permutation Testing


IEEE Signal Processing Society (SPS)

Human Brain Mapping (2010)

Other links

Contribution to Open Source

Which contains:

1- Armitage’s Trend Test, usually used in genome-wide association studies (GWAS)
2- Finding SNP markers contingency tables 
3- Fast generation of Zernike radial polynomials via q-recursion

4- Fast generation of pseudo Zernike radial polynomials via p-recursion

5- Kintner's method to generate Zernike radial polynomials

6- Prata's method  to generate Zernike radial polynomials

7-  2D, 3D, ND Trapezoidal integrals

Matlab versus C/C++ language, which one is faster?

To test for the speed issue, and how fast each environment is, I computed the so called pseudo Zernike radial polynomials for 1 milion r values, up to the order 99, using an efficient method widely known as the p-recursive.

Running the C++ executable in release mode:

[alrawi@srv Release]$Pzmp                                              
Number of r values is: 1000,000 (1M)
execution time (s) of Pseudo Zernike Polynomials values is: 3.25 seconds
[alrawi@srv Release]$

As for matlab, the execution for the same size problem, by using vectorized technique, was nearly twice as that of the C function.
>> tic; R = pseudo_zernike_polynomials_p_recursive(99,0, x); toc
Elapsed time is 6.589477 seconds.

Result: For the above task, C is twice faster than matlab.



Articles in international journals listed in the ISI

Other articles in journals

Chapters in books

Articles in conference proceedings


Visitors map.