Ayoub is a PhD student in computer science since December 2010 to present. He has conducted his doctoral studies at the Mohammed V University, Faculty of Science, Rabat, Morocco, and the French national institute of research on computer sciences (INRIA), inside GEOSTAT team under the helpful supervision of Pr. D. Aboutajdine, Dr. K. Daoudi and Pr. H. Yahia. He holds a Computer Science M.Sc. degree in telecommunication and digital signal processing (Mohammed V University, Faculty of Science, Rabat, Morocco, 2010).
Major contributions and research
The thematic of his PhD thesis, titled as "Detection of Moroccan coastal upwelling using the Sea Surface Temperature (SST) Satellite Images", concerns the study of the SST patterns and ocean dynamics in Sea Surface Temperature satellite images. In fact, the mesoscale structures are delimited by thermal fronts that occur between two water masses of different temperatures; they may be often associated to a convergence zone, which is shown to contain loci of high biological activity. As a consequence, the knowledge of thermal fronts is an important factor for the fish management and aggregation. Initially, standard methods were combined and used to automatically extract and study the upwelling activity along the Moroccan Atlantic coast Tamim et al.,2014, Tamim et al.,2013. The latter have been evaluated and validated over large databases of SST images, and shown to be promising and reliable for a wide variety of oceanographic conditions. More importantly, the quality obtained by these very simple and well-known methods suggest that they can be used effectively for the analysis of oceanic features in infrared satellite imagery. However, the foremost shortcoming encountered in the preceding approaches is that they have been constructed for use other than the oceanographic context, and consequently they often work poorly when used to detect the mesoscale structures in ocean satellite images. In fact, the conventional methods used in image processing often rely on simple hypotheses, compared to the physical of the phenomena encountered. In particular, an SST image is a measure of a thermodynamical variable, the temperature of the ocean’s microscopic upper layer in a turbulent flow (the surface of the ocean); the existence of turbulence implies the formation of extremely complex structures that conventional (linear) approaches can’t apprehend, mainly because of the absence of localization in the Fourier transform. As a matter of fact, the intermittency and the existence of multiscale organization present in Fully Developed Turbulence (FDT) make the upwelling region a complex system relevant to nonlinear signal processing. In fact, transitions of the graylevel values inside complex natural images, such that SST images, are related to the existence of a hierarchy of multiscale structures characteristic of FDT. Singularities in SST images corresponding to sets of pixels with sharp and strong variations in intensities values, unlock the multiscale organization. In this sense a novel formalism called the Microcanonical Multifractal Formalism (MMF) for the analysis of these singularities, seems to be very interesting for our problem, which is based on the strength of variations around any given pixel, which yields to a precise description of upwelling fronts in SST images.
Research interests
Image segmentation and classification applied to remotely sensed oceanographic data
Non-linear signal processing applied to infrared satellite data
Academic Honors
He obtained the best Paper Award in the Doctoral day on Information Technology and communication (JDTIC), Rabat, Morocco, 2014
He won the First prize in the National Competition of Robotic, Mohammed V University, Faculty of Science, Rabat, Morocco, 2013
