Basic Informations
C.V
Date of Birth: 25th of April 1987
Nationality: Egyptian
Google Scholar: https://scholar.google.com/Hossam Zawbaa
ResearchGate: https://www.researchgate.net/pro le/Hossam Zawbaa
LinkedIn: http://www.linkedin.com/in/hossamzawbaa
Website: http://www.cs.ubbcluj.ro/hossam.zawbaa
Military Status: Completed
I have authored/co-authored over 60 publications in peer-reviewed reputed journals and international conference proceedings. As per Google Scholar, I have more than 1000 citations, 19 h-index, and 31 i10-index as well. As per ResearchGate, I have more than 37 impact points and more than 46,500 readings. My research interests are in the area of Computational Intelligence, Machine Learning, Computer Vision,
Image Processing, and Pattern Recognition. They include both theoretical and algorithmic improvement as well as applications for various problems, such as classi cation, regression, clustering, and data mining.
- Faculty of Mathematics and Computer Science, Babes-Bolyai University, Romania.
- PhD, Machine Learning, July 2018
- Thesis Topic: Computational Intelligence Modeling of Pharmaceutical Roll Compaction
2. Faculty of Computers and Information, Cairo University, Egypt.
- MSc, Image Processing, May 2012
- Thesis Topic: Automatic Soccer Video Summarization
- BSc, Computers and Information, July 2008
Master Title
Automatic Soccer Video Summarization
Master Abstract
Soccer is one of the most popular team's sports all over the world. Most sports games are naturally organized into successive and alternating plays of offense and defense, cumulating at events such as goal or attack. If the sports videos can be segmented according to these semantically meaningful events, it then can be used in numerous applications to enhance their values and enrich the user's viewing experiences. According to this, soccer video summarization and analysis has recently attracted much research and a wide spectrum of possible applications have been considered.
Soccer video summarization and analysis is concerned with the extraction of valuable semantics by efficient and effective processing of a combination of visual, audio and text information. However, one of the major limitations of current soccer analysis is the semantic gap between the low-level features such as (color, texture, shape and motion) and high-level representation such as (shot types, shot length, and shot replays).\\
This thesis presents an automatic soccer video summarization system using machine learning (ML) techniques. The proposed system is composed of five phases. Namely; in the pre-processing phase, the system segments the whole video stream into small video shots. Then, in the shot processing phase, it applies two types of classification (shot type classification and play / break classification) to the video shots resulted from the pre-processing phase. Afterwards, in the replay detection phase, the proposed system applies two machine learning algorithms, namely; support vector machine (SVM) and artificial neural network (ANN), for emphasizing important segments with championship logo appearance. Also, in the excitement event detection phase, the proposed system uses both machine learning algorithms for detecting the scoreboard which contain an information about the score of the game. The proposed system also uses k-means algorithm and Hough line transform for detecting vertical goal posts and Gabor filter for detecting goal net. Finally, in the event detection and summarization phase, the proposed system highlights the most important events during the match. Experiments on real soccer videos demonstrate encouraging results. The event detection and summarization has attained recall 94\% and precision 97.3\% for soccer match videos from five international soccer championships.
PHD Title
Computational Intelligence Modeling of Pharmaceutical Roll Compaction
PHD Abstract
Datasets ordinarily include a huge number of features (attributes) with irrelevant and redundant features. Selecting a subset of relevant properties from a large set of features that describe a dataset is a challenging machine learning task. Feature selection algorithms explore the data to eliminate noisy, irrelevant, redundant data, and simultaneously optimize the classification performance. In biology, for instance, the advances in the available technologies enable the generation of an enormous number of biomarkers that describe the data. Selecting the more relevant biomarkers along with a very high prediction performance over the data can be a daunting task, particularly if the data are high-dimensional. \emph{Feature selection} is formulated as a multi-objective optimization problem with the objective of maximizing the performance of the prediction model while minimizing the number of the selected features.
The search space size is increasing exponentially with respect to the number of features in a dataset. In practice, searching for the optimal solution is not feasible for even medium-sized datasets. A diversity of stochastic techniques gives promising feature selection procedures such as genetic algorithms (GA), particle swarm optimization (PSO), and much more. In the best case, the selected features will improve the prediction performance and provide a faster and more cost-effective prediction that leads to comparable or even better performance than using the full feature set. Most of the recent optimization algorithms are inspired by nature and its main sources like \emph{biology}, \emph{physics}, or \emph{chemistry}. The optimization algorithms based on \emph{biological} field are called \emph{bio-inspired}.
The balance between exploration of the search space and exploitation of the best solutions is a challenge in multi-objective optimization. A good optimization algorithm needs to adjust the exploration and exploitation adaptively to find the optimal solution quickly. The proposed bio-inspired optimization algorithms are applied in wrapper-based feature selection mode to select the optimal feature subset that maximizes classification accuracy (minimizes prediction error) while reducing the number of selected features. In wrapper-based feature selection, a machine learning technique is used in the evaluation step. Despite the fact that it is costly in computational time, this technique has a good prediction performance.
In this study, we have been developed new variants of the native bio-inspired optimization algorithms such as binary antlion optimization (BALO), chaotic antlion optimization (CALO), L\`{e}vy antlion optimization (LALO), binary grey wolf optimization (BGWO), and much more. All the proposed optimization algorithms were compared to two well-known algorithms used in feature selection, namely particle swarm optimization (PSO) and genetic algorithm (GA). A set of assessment indicators were used to evaluate generalization ability and prediction performance of the algorithms in the classification problem over 21 datasets from the UCI repository (10 datasets in the regression problem). The experimental results prove the capability of the proposed variants of the native optimization algorithms to explore the search space for the optimal feature subset regardless of the initialization methods and the used stochastic operators.
In the pharmaceutical industry, to develop new formulations or products and to optimize manufacturing processes are often used the exploitation of knowledge on the causal relationship between product quality and attributes of formulations. With the big data captured in the pharmaceutical product development practice, computational intelligence (CI) models, based on machine learning and bio-inspired optimization algorithms, could potentially be used to identify critical quality attributes (CQA) and critical process parameters (CPP), for the formulations and manufacturing processes. The principal objective of our study in the pharmaceutical field is to evaluate the robustness of machine learning techniques combined with bio-inspired optimization algorithms in modeling tablet manufacturing processes. More precisely, our effort is focused on the prediction of tablet properties such as porosity and tensile strength from powder and ribbons characteristics. For this purpose, roll compaction experiments were performed with various pharmaceutical excipients (MCC PH 101, MCC PH 102, MCC DG, Mannitol Pearlitol 200SD, Lactose, and binary mixtures), leading to datasets with a large number of attributes (features). The modeling efficiency is evaluated regarding the average of selected features size (reduction) and the root mean square error (RMSE). We have remarked that the predicted results were in good agreement with the actual experimental data.