Zainab Ouardirhi

ARES (2021 - 2023)

Biography

Zainab Ouardirhi pursued her academic studies at the Faculty of Sciences in Rabat, Mohammed V University. She earned her Bachelor's degree in "Mathematical Sciences and Computer Science," during which she undertook a final project in the field of artificial intelligence. Subsequently, she completed her Master's degree in "Applied Offshoring Computing," where her research project was titled "Intelligent System for Moroccan License Plate Detection and Recognition" conducted at the Moroccan Foundation for Advanced Sciences, Innovation, and Scientific Research (MAScIR).

Zainab Ouardirhi is a Deep Learning Nanodegree Graduate. Currently, she is actively engaged in research within the realm of deep learning, with a particular emphasis on computer vision for video surveillance. This research is conducted in collaboration between the Computer Science, Software, and Artificial Intelligence (ILIA) department at the Faculty of Engineering of Mons, University of Mons, Belgium, and the Smart Systems Lab (SSLab) at the National School of Computer Science and Systems Analysis, Mohammed V University, Morocco.

In her research, Zainab Ouardirhi is dedicated to advancing the field of deep learning, with a specific focus on applications related to computer vision in video surveillance. This work contributes to bridging the gap between academia and real-world challenges, ultimately seeking to enhance the capabilities of intelligent systems for video analysis and surveillance applications.

Thesis

Incorporating AI and Deep Learning for Robust 2D/3D Object Detection in the Presence of Occlusion

ABSTRACT :

In this thesis, we focus on developing a self-evolving learning approach for robust 2D/3D object detection, even in the presence of occlusions and data distortions caused by weather changes. Our methodology combines supervised and unsupervised learning methods to reduce annotation efforts, minimize human intervention, discover new features, and enhance model performance. We also emphasize data preprocessing to address data-related issues

This research aims to tackle the challenges of object detection in complex environments by leveraging advanced AI and deep learning techniques. Our approach enables the system to adapt and perform well under challenging conditions, contributing to the advancement of object detection in practical scenarios.

KEYWORDS : Deep Learning, Occlusion Handling, Data Preprocessing, Self-evolving Learning

Advisors

Sidi Ahmed Mahmoudi

Embedded and Explainable AI

Check Profile

Mohammed Benjelloun

Mostaph Zbakh

Publications
Enhancing Object Detection in Smart Video Surveillance: A Survey of Occlusion-Handling Approaches

Electronics 13 (3), 541, 2024

Access publication
An Efficient real-time Moroccan automatic license plate recognition system based on the YOLO object detector

International Conference On Big Data and Internet of Things, 290-302, 2022

Access publication
FuDensityNet: Fusion-Based Density-Enhanced Network for Occlusion Handling

Proceedings Copyright 632, 639, 2024

Access publication