Mohamed Chtourou

Facial expression recognition based on a mlp neural network using constructive training algorithm

This paper presents a constructive training algorithm for Multi Layer Perceptron (MLP) applied to facial expression recognition applications. The developed algorithm is composed by a single hidden-layer using a given number of neurons and a small number of training patterns. When the Mean Square Error MSE on the Training Data TD is not reduced to a predefined value, the number of hidden neurons grows during the neural network learning. Input patterns are trained incrementally until all patterns of TD are presented and learned. The proposed MLP constructive training algorithm seeks to find synthesis parameters as the number of patterns corresponding for subsets of each class to be presented initially in the training step, the initial number of hidden neurons, the number of iterations during the training step as well as the MSE predefined value. The suggested algorithm is developed in order to classify a facial expression. For the feature extraction stage, a biological vision-based facial description, namely Perceived Facial Images PFI has been applied to extract features from human face images. To evaluate, the proposed approach is tested on three databases which are the GEMEP FERA 2011, the Cohn-Kanade facial expression and the facial expression recognition FER-2013 databases. Compared to the fixed MLP architecture and the literature review, experimental results clearly demonstrate the efficiency of the proposed algorithm.

A hybrid approach for training recurrent neural networks: application to multi-step-ahead prediction of noisy and large data sets

Noisy and large data sets are extremely difficult to handle and especially to predict. Time series prediction is a problem, which is frequently addressed by researchers in many engineering fields. This paper presents a hybrid approach to handle a large and noisy data set. In fact, a Self Organizing Map (SOM), combined with multiple recurrent neural networks (RNN) has been trained to predict the components of noisy and large data set. The SOM has been developed to construct incrementally a set of clusters. Each cluster has been represented by a subset of data used to train a recurrent neural network. The back propagation through time has been deployed to train the set of recurrent neural networks. To show the performances of the proposed approach, a problem of instruction addresses prefetching has been treated.

A new method for fuzzy rule base reduction

This paper presents a new approach for fuzzy rule base reduction using similarity concepts and interpolation techniques. The algorithm consists on: First, measure similarity between rules for the best choice of which of them will be deleted. This operation is done without modification of membership functions. Second, if a new input data is presented to the fuzzy system, interpolation techniques will be used to take into account this arriving data. The main idea of this work is to improve accuracy of the fuzzy system after reduction step. A comparative study between three interpolation methods is done. A mathematical case is treated to show the performance of the proposed method.

Neural network control of nonlinear dynamic systems using hybrid algorithm

A hybrid method is proposed to control a nonlinear dynamic system.This hybrid algorithm combines gradient method and Kohonen algorithm to obtain faster convergence.The proposed algorithm can considerably reduce networking time. In this paper, a hybrid method is proposed to control a nonlinear dynamic system using feedforward neural network. This learning procedure uses different learning algorithm separately. The weights connecting the input and hidden layers are firstly adjusted by a self organized learning procedure, whereas the weights between hidden and output layers are trained by supervised learning algorithm, such as a gradient descent method. A comparison with backpropagation (BP) shows that the new algorithm can considerably reduce network training time.

Face recognition based on perceived facial images and multilayer perceptron neural network using constructive training algorithm

This study presents a modified constructive training algorithm for multilayer perceptron (MLP) which is applied to face recognition problem. An incremental training procedure has been employed where the training patterns are learned incrementally. This algorithm starts with a small number of training patterns and a single hidden-layer using an initial number of neurons. During the training, the hidden neurons number is increased when the mean square error (MSE) threshold of the training data (TD) is not reduced to a predefined value. Input patterns are trained incrementally until all patterns of TD are learned. The aim of this algorithm is to determine the adequate initial number of hidden neurons, the suitable number of training patterns in the subsets of each class and the number of iterations during the training step as well as the MSE threshold value. The proposed algorithm is applied in the classification stage in face recognition system. For the feature extraction stage, this paper proposes to use a biological vision-based facial description, namely perceived facial images, applied to extract features from human face images. Gabor features and Zernike moment have been used in order to determine the best feature extractor. The proposed approach is tested on the Cohn-Kanade Facial Expression Database. Experimental results indicate that a good architecture of neural network classifier can be obtained. The effectiveness of the proposed method compared with the fixed MLP architecture has been proved.

Multilayer perceptron training using an evolutionary algorithm

It is shown through a considerably large literature review that combinations of Artificial Neural Networks (ANNs) and Evolutionary Algorithms (EAs) can lead to significantly better intelligent systems than relying on ANNs or EAs alone. Evolution can be introduced into ANNs at many different levels. This paper focuses on the evolution of connection weights, which provides a global approach to connection weight training especially when gradient information of the error function is difficult or costly obtained. Due to the simplicity and generality of the evolution and the fact that gradient-based training algorithm often have to be run multiple times in order to avoid being trapped in a poor local optimum, the evolutionary approach is quite competitive. This paper takes a step in that direction by introducing an EA for Multi-Layer Perceptron (MLP) learning, called Perceptron Learning using Genetic algorithm (PLG), that gets results comparably better than BackPropagation (BP).

A learning-automaton-based method for fuzzy inference system identification

Abstract A learning method of fuzzy inference rules using learning automata is described. The tuning of the membership functions in the antecedent part and the real numbers in the consequent part of the inference rule can be stated as an optimization problem. Learning automata have been used for control and optimization purposes. The optimization algorithm is based on a hierarchical structure of stochastic automata with variable structures. Results related to numerical examples and fuzzy modelling are presented. Simulation results show the performance and the implementation simplicity of the proposed method.

Facial Expression Recognition Based on Perceived Facial Images and Local Feature Matching

Facial expression recognition is to determine the emotional state of the face regardless of its identity. Deriving an effective facial representation from original face images is a vital step for successful facial expression recognition. This paper presents a biological vision-based facial description, called Perceived Facial Images “PFI” applied to facial expression recognition. For the classification step, Scale Invariant Feature Transform “SIFT” is used to extract a local feature in images. Then, a matching computation is processed between a testing image and all train images for recognizing facial expression. To evaluate, the proposed approach is tested on the GEMEP FERA 2011 database and the Cohn-Kanade Facial Expression database. To compare, the developed algorithm achieves better experimental results than the other approaches in the literature.

Optimised fuzzy logic controller for a mobile robot navigation

This paper presents fuzzy logic controller (FLC) design using optimisation techniques. The FLC has been developed and implemented for the motion of the robot from an initial position towards another desired position, taking into account the kinematic constraints. First, we have carried out a simulation of a fuzzy logic based controller which determines the speed values of each driving wheel, while the robot seeking the goal. Second, an optimisation of this controller has been realised using gradient method or genetic algorithms. Simulation and experimental results demonstrates the effectiveness of the proposed approach.

Polyphenolic compounds progress during olive mill wastewater sludge and poultry manure co-composting, and humic substances building (Southeastern Tunisia)

In Mediterranean areas, olive mill wastes pose a major environmental problem owing to their important production and their high polyphenolic compounds and organic acids concentrations. In this work, the evolution of polyphenolic compounds was studied during co-composting of olive mill wastewater sludge and poultry manure, based on qualitative (G-50 sephadex) and quantitative (Folin–Ciocalteu), as well as high pressure liquid chromatography analyses. Results showed a significant polyphenolic content decrease of 99% and a noticeable transformation of low to high molecular weight fraction during the compost maturation period. During this step, polyphenols disappearance suggested their assimilation by thermophilic bacteria as a carbon and energy source, and contributed to humic substances synthesis. Polyphenolic compounds, identified initially by high pressure liquid chromatography, disappeared by composting and only traces of caffeic, coumaric and ferulic acids were detected in the compost. In the soil, the produced compost application improved the chemical and physico–chemical soil properties, mainly fertilising elements such as calcium, magnesium, nitrogen, potassium and phosphorus. Consequently, a higher potato production was harvested in comparison with manure amendment.