Papers by Mohammed El Haj Tirari
Australian & New Zealand Journal of Statistics, 2003
SummaryThe Montanari (1987) regression estimator is optimal when the population regression coeffi... more SummaryThe Montanari (1987) regression estimator is optimal when the population regression coefficients are known. When the coefficients are estimated, the Montanari estimator is not optimal and can be extremely volatile. Using design‐based arguments, this paper proposes a simpler and better alternative to the Montanari estimator that is also optimal when the population regression coefficients are known. Moreover, it can be easily implemented as it involves standard weighted least squares. The estimator is applicable under single stage stratified sampling with unequal probabilities within each stratum.
A new optimal quadratic predictor of residual linear model in a finite population
International Journal of Advanced Computer Science and Applications
No one can deny that the deaf-mute community has communication problems in daily life. Advances i... more No one can deny that the deaf-mute community has communication problems in daily life. Advances in artificial intelligence over the past few years have broken through this communication barrier. The principal objective of this work is creating an Arabic Sign Language Recognition system (ArSLR) for recognizing Arabic letters. The ArSLR system is developed using our image pre-processing method to extract the exact position of the hand and we proposed architecture of the Deep Convolutional Neural Network (CNN) using depth data. The goal is to make it easier for people who have hearing problems to interact with normal people. Based on user input, our method will detect and recognize hand-sign letters of the Arabic alphabet automatically. The suggested model is anticipated to deliver encouraging results in the recognition of Arabic sign language with an accuracy score of 97,07%. We conducted a comparison study in order to evaluate proposed system, the obtained results demonstrated that this method is able to recognize static signs with greater accuracy than the accuracy obtained by similar other studies on the same dataset used.
Trajectories Modeling and Clustering
Proceedings of the International Conference on Learning and Optimization Algorithms: Theory and Applications
1 The location of moving objects facilitates the monitoring of their evolution and the history of... more 1 The location of moving objects facilitates the monitoring of their evolution and the history of displacement offers interesting perspectives in the field of the study of the behavior of these objects. The purpose of this paper is to provide a new methodology for constructing object trajectories based on a matching process that implements different features such as occultation management, surface, histogram, and so on. a second objective was to propose a classification approach of its trajectories. The proposed algorithm consists in classifying its trajectories in homogenous groups and isolating aberrant trajectories that exhibit deviant behavior. Its principle is based on the transformation of trajectories into spaces of distinct characteristics. These spaces make it possible to provide additional information on the characteristics of the model of movement of an object in a video sequence. The approach then consists in using the Mean-Shift algorithm to estimate the groups in each characteristic space. Clusters with the small number of associated trajectories and trajectories that are far from the center of the groups are considered outliers.
IAES International Journal of Artificial Intelligence (IJ-AI)
Facial expression recognition (FER) represents one of the most prevalent forms of interpersonal c... more Facial expression recognition (FER) represents one of the most prevalent forms of interpersonal communication, which contains rich emotional information. But it became even more challenging during the times of COVID, where face masks became a mandatory protection measure, leading to the challenge of occluded lower-face during facial expression recognition. In this study, deep convolutional neural network (DCNN) represents the core of both our full-face FER system and our masked face FER model. The focus was on incorporating knowledge distillation in transfer learning between a teacher model, which is the full-face FER DCNN, and the student model, which is the masked face FER DCNN via the combination of both the loss from the teacher soft-labels vs the student soft labels and the loss from the dataset hard-labels vs the student hard-labels. The teacher-student architecture used FER2013 and a masked customized version of FER2013 as datasets to generate an accuracy of 69% and 61% respe...
Enhancing cash management using machine learning
2019 1st International Conference on Smart Systems and Data Science (ICSSD)
Cash management is a complicated task since it is the interaction between multiple monetary activ... more Cash management is a complicated task since it is the interaction between multiple monetary activities including collections, disbursements, concentration, investments and funding [1], moreover, it can be easily influenced by several unpredictable internal and external factors from different areas. A misuse or underestimation may lead to devastating financial consequences. To manage the cash requires painstaking approaches, to approximate its size requires advanced and meticulous tools.By relying on machine learning concepts, we attempt to build an intelligent tool adapted to the public expenditure management sector. Giving a set of payment orders in progress, the model is conceived to allow the cash managers to predict the amounts to be drawn in a period of time, thus, to have a clear vision over cash trend.The experiments demonstrate the applicability of the model and exhibit encouraging prediction results. Yet, we believe that still there are unexplored features to be considered and leveraged to enhance the model performance and specially the accuracy which is valuable for a crucial financial decision.
This paper we propose a simple and efficient video surveillance system which detects and tracks a... more This paper we propose a simple and efficient video surveillance system which detects and tracks a person in a video stream. Object detection is the most important and crucial step for any video surveillance system. In this paper, we separate foreground and background by using the statistical model of Gaussian Mixture (GMM). Interests points are identified in the detected regions (foreground) using the Harris detector. The analysis of connected components detected by subtracting the background allows grouped the pixels of moving objects in order to extract the center of gravity. Then, a boundary box is used to limit the area of connected components in order to detect the coordinates of the gravity center of the moving person. Finally, by projection in Euclidean plan we can get the trajectory of the person in motion and compute the Euclidean distance crossed in the scene.
Prediction Demand for Classified Ads Using Machine Learning
Proceedings of the 2nd International Conference on Networking, Information Systems & Security - NISS19, 2019
Classified ads prediction is a very interesting activity for organizations in order to increase t... more Classified ads prediction is a very interesting activity for organizations in order to increase the purchase quantity of a product and thereafter the possibility of sale. Used goods predicting can be done by calculating the probability of sale for each selected product. In this paper, we conduct an empirical analysis on classified ads prediction of Avito dataset in order to develop prediction models using three individual machine-learning techniques and five ensemble learners. We compare and evaluate the performance of the proposed models using Root Mean Square Error (RMSE) measure. The stacked generalization method was also used to combine the best-performed models to select the best one. The results show that the Extreme Gradient Boosting Machine algorithm (XGBoost) is the most accurate model with an RMSE value of 0.2253.
Video Surveillance: Analyzing People’s Movements in a Closed Environment
International Review on Computers and Software, Mar 31, 2014
In this paper, we developed a method for tracking multiple people in a closed environment based o... more In this paper, we developed a method for tracking multiple people in a closed environment based on the feed of a surveillance camera. Monitoring the evolution of the position of people in real time has enabled us to get information on areas occupied by each person in a scene. In addition, to consolidate the trajectories obtained in homogeneous classes, we proposed a new algorithm which allows to adapt the hierarchical classification technique (CHA) and the K-Means technique in cases of similarity measures between trajectories
Pratiques et méthodes de sondage
Dunod eBooks, 2011
The detection of moving objects in a video sequence is an essential step in almost all the system... more The detection of moving objects in a video sequence is an essential step in almost all the systems of vision by computer. However, because of the dynamic change in natural scenes, the detection of movement becomes a more difficult task. In this work, we propose a new method for the detection moving objects that is robust to shadows, noise and illumination changes. For this purpose, the detection phase of the proposed method is an adaptation of the MOG approach where the foreground is extracted by considering the HSV color space. To allow the method not to take shadows into consideration during the detection process, we developed a new shade removal technique based on a dynamic thresholding of detected pixels of the foreground. The calculation model of the threshold is established by two statistical analysis tools that take into account the degree of the shadow in the scene and the robustness to noise. Experiments undertaken on a set of video sequences showed that the method put forward provides better results compared to existing methods that are limited to using static thresholds.
Uploads
Papers by Mohammed El Haj Tirari