MOHAMMED OUANAN - Academia.edu (original) (raw)
Uploads
Papers by MOHAMMED OUANAN
Abstract. This paper concerns some special properties of the principal eigen-value of nonlinear e... more Abstract. This paper concerns some special properties of the principal eigen-value of nonlinear elliptic systems with Dirichlet boundary conditions. We study the stability with respect to the exponents p and q; and the dependence on the domain variations. 1.
Electronic Journal of Qualitative Theory of Differential Equations, 2004
Electronic Journal of Qualitative Theory of Differential Equations, 2003
Applicationes Mathematicae, 2009
Nonlinear Differential Equations and Applications NoDEA, 2008
Canadian Mathematical Bulletin, 2006
We show that each point of the principal eigencurve of the nonlinear problem–Δpu– λm(x)|u|p–2u= μ... more We show that each point of the principal eigencurve of the nonlinear problem–Δpu– λm(x)|u|p–2u= μ|u|p–2uin Ω,is stable (continuous) with respect to the exponentpvarying in (1,∞); we also prove some convergence results of the principal eigenfunctions corresponding.
The multilayer perceptron (MLP) is an artificial neural network composed of one or more hidden la... more The multilayer perceptron (MLP) is an artificial neural network composed of one or more hidden layers. MLP has an extensive array of classification and regression applications in a wide range of domains. Selecting the right neural network architecture to solve a specific problem is one of the major areas of neural network research. The number of neurons in the hidden layers and the initial weights have a great effect on the convergence of MLP training algorithms. This paper presents a technique termed hyperparameter optimization model for the multilayer perceptron. We model in terms of mixed-variable optimization problems with nonlinear constraints. To solve the obtained model, a hybrid algorithm is used, based on the genetic algorithm and the backpropagation algorithm. The numerical results demonstrate the efficiency of the technique presented in this paper, and the advantages of the proposed model vis-a-vis the previous state-of-the-art model.
Lecture notes in networks and systems, 2023
Procedia Computer Science, 2017
TELKOMNIKA : Indonesian Journal of Electrical Engineering, Aug 1, 2015
Advances in intelligent systems and computing, 2022
Face landmarking, defined as the detection and localization of certain keypoints points on the fa... more Face landmarking, defined as the detection and localization of certain keypoints points on the face, plays arguably the important role as an intermediary step for many subsequent face processing operations that ranges from biometric recognition to the understanding of mental states. Though its conceptual simplicity, the computer vision problem has proven extremely challenging due to multitude of compound factors such as pose, expression, occlusion and illumination. In this paper, we survey the recent advances of facial landmarks localization techniques and discuss the advantages and disadvantages of the presented algorithms.
International Journal of Advanced Computer Science and Applications, 2023
in this present work, we emphasize the main question about success: why do some people seem to be... more in this present work, we emphasize the main question about success: why do some people seem to be successful while the great majority of the rest of us seem not to be? To be more specific, the challenge of this work is to propose a novel system to assist people in developing good habits using deep learning techniques. To achieve this goal, we propose the use of Recurrent Neural Networks (RNN) and Convolutional Neural Networks (CNN). The two best-known supervised learning method. The first step consists of using a deep RNN neural network to perform the captioning of the user’s activities. Then use another deep CNN to do the classification of the activity based on the result of the first network and then decide whether what the user is doing is a good habit or not. To our knowledge, there is no previous work dealing with this topic through a computer science lens, and this will be of greater value to most of the people who are interested in developing new good habits. This system will also suggest new directions for basic and applied research on success in general and good habits in particular.
Today, we witness the appearance of many lifelogging cameras that are able to capture the life of... more Today, we witness the appearance of many lifelogging cameras that are able to capture the life of a person wearing the camera, which produce a large number of images everyday. Automatically characterizing the experience and extracting patterns of behavior of individuals from this huge collection of unlabeled and unstructured egocentric data present major challenges and require novel and efficient algorithmic solutions. The main goal of this work is to propose a new method to automatically assess day similarity from the lifelogging images of a person. We propose a technique to measure the similarity between images based on the Swain's distance and generalize it to detect similarity between daily visual data. To this purpose, we apply dynamic time warping (DTW) combined with the Swain's distance for final day similarity estimation. For validation, we apply our technique on the Egocentric Dataset of University of Barcelona (EDUB) of 4912 daily images acquired by 4 persons with preliminary encouraging results.
Lecture notes in networks and systems, Nov 12, 2017
Face verification in unconstrained images, remains a challenging problem. Many works have been pr... more Face verification in unconstrained images, remains a challenging problem. Many works have been proposed to solve this problem. However, the performance gap existing between the human visual system and machines in face recognition remain important. This paper makes two contributions: firstly, for improving face recognition in the wild, at least in terms of pose variations, we propose a method for aligning faces by employing single-3D face model as reference produced by FaceGen Modeller. Secondly, we developed a novel face representation technique based on Gabor Filters. The proposed approach relies on combination of Gabor magnitude and Gabor phase informations into an unified framework, which capable to surpass standard representations in the well-known FERET dataset.
Most retrieval systems are mostly based on key words for image search, but in many case it cannot... more Most retrieval systems are mostly based on key words for image search, but in many case it cannot meet demands for different user with different view. In this paper, a new approach of content based image retrieval (CBIR) is presented, which is based on the image frequency content. Indeed, we have used the 2-D ESPRIT (Estimation of Signal Parameters via Rotationnal Invariance Techniques) method to extract from the image the frequency content for constructing the vector descriptor. Our approach is applied to the Coil_100 database and the experimental results show that this method improves the precision of image retrieval.
2017 International Conference on Wireless Technologies, Embedded and Intelligent Systems (WITS)
With the appearance of many devices that everyday captured a large number of images. The rapid ac... more With the appearance of many devices that everyday captured a large number of images. The rapid access to these huge collections of images and automatically characterizing an activity or an experience from this huge collection of unlabeled and unstructured egocentric data presents major challenges and requires novel and efficient algorithmic solutions. One of the big challenges of egocentric vision and lifelogging is to develop automatic algorithms to automatically characterize everyday activities. Such information is of high interest to predict migraines attacks or assure healthy behavior of patients and individuals of high healthy risk. In this work, we first conduct a comprehensive survey of existing egocentric datasets and we will present our future contribution to automatically characterize everyday activities.
Digital Technologies and Applications
Abstract. This paper concerns some special properties of the principal eigen-value of nonlinear e... more Abstract. This paper concerns some special properties of the principal eigen-value of nonlinear elliptic systems with Dirichlet boundary conditions. We study the stability with respect to the exponents p and q; and the dependence on the domain variations. 1.
Electronic Journal of Qualitative Theory of Differential Equations, 2004
Electronic Journal of Qualitative Theory of Differential Equations, 2003
Applicationes Mathematicae, 2009
Nonlinear Differential Equations and Applications NoDEA, 2008
Canadian Mathematical Bulletin, 2006
We show that each point of the principal eigencurve of the nonlinear problem–Δpu– λm(x)|u|p–2u= μ... more We show that each point of the principal eigencurve of the nonlinear problem–Δpu– λm(x)|u|p–2u= μ|u|p–2uin Ω,is stable (continuous) with respect to the exponentpvarying in (1,∞); we also prove some convergence results of the principal eigenfunctions corresponding.
The multilayer perceptron (MLP) is an artificial neural network composed of one or more hidden la... more The multilayer perceptron (MLP) is an artificial neural network composed of one or more hidden layers. MLP has an extensive array of classification and regression applications in a wide range of domains. Selecting the right neural network architecture to solve a specific problem is one of the major areas of neural network research. The number of neurons in the hidden layers and the initial weights have a great effect on the convergence of MLP training algorithms. This paper presents a technique termed hyperparameter optimization model for the multilayer perceptron. We model in terms of mixed-variable optimization problems with nonlinear constraints. To solve the obtained model, a hybrid algorithm is used, based on the genetic algorithm and the backpropagation algorithm. The numerical results demonstrate the efficiency of the technique presented in this paper, and the advantages of the proposed model vis-a-vis the previous state-of-the-art model.
Lecture notes in networks and systems, 2023
Procedia Computer Science, 2017
TELKOMNIKA : Indonesian Journal of Electrical Engineering, Aug 1, 2015
Advances in intelligent systems and computing, 2022
Face landmarking, defined as the detection and localization of certain keypoints points on the fa... more Face landmarking, defined as the detection and localization of certain keypoints points on the face, plays arguably the important role as an intermediary step for many subsequent face processing operations that ranges from biometric recognition to the understanding of mental states. Though its conceptual simplicity, the computer vision problem has proven extremely challenging due to multitude of compound factors such as pose, expression, occlusion and illumination. In this paper, we survey the recent advances of facial landmarks localization techniques and discuss the advantages and disadvantages of the presented algorithms.
International Journal of Advanced Computer Science and Applications, 2023
in this present work, we emphasize the main question about success: why do some people seem to be... more in this present work, we emphasize the main question about success: why do some people seem to be successful while the great majority of the rest of us seem not to be? To be more specific, the challenge of this work is to propose a novel system to assist people in developing good habits using deep learning techniques. To achieve this goal, we propose the use of Recurrent Neural Networks (RNN) and Convolutional Neural Networks (CNN). The two best-known supervised learning method. The first step consists of using a deep RNN neural network to perform the captioning of the user’s activities. Then use another deep CNN to do the classification of the activity based on the result of the first network and then decide whether what the user is doing is a good habit or not. To our knowledge, there is no previous work dealing with this topic through a computer science lens, and this will be of greater value to most of the people who are interested in developing new good habits. This system will also suggest new directions for basic and applied research on success in general and good habits in particular.
Today, we witness the appearance of many lifelogging cameras that are able to capture the life of... more Today, we witness the appearance of many lifelogging cameras that are able to capture the life of a person wearing the camera, which produce a large number of images everyday. Automatically characterizing the experience and extracting patterns of behavior of individuals from this huge collection of unlabeled and unstructured egocentric data present major challenges and require novel and efficient algorithmic solutions. The main goal of this work is to propose a new method to automatically assess day similarity from the lifelogging images of a person. We propose a technique to measure the similarity between images based on the Swain's distance and generalize it to detect similarity between daily visual data. To this purpose, we apply dynamic time warping (DTW) combined with the Swain's distance for final day similarity estimation. For validation, we apply our technique on the Egocentric Dataset of University of Barcelona (EDUB) of 4912 daily images acquired by 4 persons with preliminary encouraging results.
Lecture notes in networks and systems, Nov 12, 2017
Face verification in unconstrained images, remains a challenging problem. Many works have been pr... more Face verification in unconstrained images, remains a challenging problem. Many works have been proposed to solve this problem. However, the performance gap existing between the human visual system and machines in face recognition remain important. This paper makes two contributions: firstly, for improving face recognition in the wild, at least in terms of pose variations, we propose a method for aligning faces by employing single-3D face model as reference produced by FaceGen Modeller. Secondly, we developed a novel face representation technique based on Gabor Filters. The proposed approach relies on combination of Gabor magnitude and Gabor phase informations into an unified framework, which capable to surpass standard representations in the well-known FERET dataset.
Most retrieval systems are mostly based on key words for image search, but in many case it cannot... more Most retrieval systems are mostly based on key words for image search, but in many case it cannot meet demands for different user with different view. In this paper, a new approach of content based image retrieval (CBIR) is presented, which is based on the image frequency content. Indeed, we have used the 2-D ESPRIT (Estimation of Signal Parameters via Rotationnal Invariance Techniques) method to extract from the image the frequency content for constructing the vector descriptor. Our approach is applied to the Coil_100 database and the experimental results show that this method improves the precision of image retrieval.
2017 International Conference on Wireless Technologies, Embedded and Intelligent Systems (WITS)
With the appearance of many devices that everyday captured a large number of images. The rapid ac... more With the appearance of many devices that everyday captured a large number of images. The rapid access to these huge collections of images and automatically characterizing an activity or an experience from this huge collection of unlabeled and unstructured egocentric data presents major challenges and requires novel and efficient algorithmic solutions. One of the big challenges of egocentric vision and lifelogging is to develop automatic algorithms to automatically characterize everyday activities. Such information is of high interest to predict migraines attacks or assure healthy behavior of patients and individuals of high healthy risk. In this work, we first conduct a comprehensive survey of existing egocentric datasets and we will present our future contribution to automatically characterize everyday activities.
Digital Technologies and Applications