03.Forum of Artificial Intelligence and Its Applications

Permanent URI for this collectionhttps://dspace.univ-eloued.dz/handle/123456789/10731

Browse

Now showing 1 - 20 of 79

Segmentation of the Breast Masses in Mammograms Using Active Contour for Medical Practice: AR based surgery
(University of Eloued جامعة الوادي, 2022-01-24) Guerroudji, Mohamed Amine; Amara, Kahina; Aouam, Djamel; Zenati, Nadia; Djekoune, Oualid; Benbelkacem, Samir
Images have been one of the most important ways humans have used to communicate and impart knowledge and information since the dawn of mankind, as an image can encompass a large amount of information concerning the quality of life linked to health and particularly in oncology precisely the breast cancer. New technologies such as Augmented Reality (AR) guidance allows a surgeon to see sub-surface structures, by overlaying pre-operative imaging data on a live laparoscopic video. The presence of masses in mammography is particularly interesting for the early detection of the breast cancer. In this article, we propose to use a mass detection system, based on two main axes: segmentation and pretreatment. The latter is based on the suppression of the noise by a Gaussian filter and mathematical morphology (white Top-Hat transform) in order to bring out all the spots (Clear Spots) possible to be pathologies. In the second axis, we are interested in the segmentation of pathologies in mammography images. This consists of segmenting the object of interest by active contour models (Chunming Li). Visually, the obtained results are very clear, and show the good performance of the new approach suggested in this work. This latter allows extracting successfully the masses starting from the mammography referents from the database Mini MIAS. The proposed breasts masses detection can, thus, provide an acceptable accuracy for an AR-based surgery or medicine courses with scene augmentation of videos, which provides a seamless use of augmented-reality for surgeons in visualizing cancer tumors.
Rational Function Model Optimization Based On Swarm Intelligence Metaheuristic Algorithms
(University of Eloued جامعة الوادي, 2022-01-24) Mezouar, Oussama; Meskine, Fatiha; Boukerch, Issam
The Rrational Function Model (RFM) is progressively being familiar to the mapping and photogrammetric researchersand has been widely used as an approximate to rigorous Models. According to it ability to preserve the complete accuracy of various types of physical sensors and its independence of sensors and platforms, it can be had with any coordination system. Nevertheless, the RFM coefficients are also known as rational polynomial coefficients (RPCs) dependent on a large number of ground control points which makes the model susceptible to over parameterization error and a time-consuming also the RPCs have no physical meaning, as a result, selecting the best combination of RPCs is difficult. The intelligent algorithms based meta-heuristic optimization seem to be an effective approach for overcoming this problem. This paper focuses on the application of recent swarm intelligence based meta-heuristic algorithms for RFM optimization. The most popular optimization methods considered are ant colony algorithm, genetic algorithms and particle swarm optimization.Furthermore in this research we proposed an parallel hybrid metaheuristic optimization algorithm that combines the genetic algorithm and particle swarm optimization concepts to overcoming the swarm intelligent limitations for RFM optimization. The different algorithms are applied for two data sets provided from the Algerian satellite (ALSAT2).The results demonstrated that the proposed method is more accurate than the threesuggested based meta-heuristic methods.
Feature Fusion for Kinship Verification based on Face Image Analysis
(University of Eloued جامعة الوادي, 2022-01-24) Zekrini, Fatima; Nemmour, Hassiba; Chibani, Youcef
This paper proposes the fusion of two new features for improving kinship verification based on face image analysis. Combined features are the Gradient Local Binary Patterns (GLBP), which associates gradient and textural information. The second descriptor is the Histogram Of Templates (HOT), which is a shape descriptor. These features are utilized with the support vector machines classifier to develop the kinship verification. Experiments are carried out on Cornell and Kinface W-II datasets. Results obtained highlight the effectiveness of the proposed system which provide competitive and sometimes better performance than the state of the art.
RONI-based medical image watermarking using DWT and LSB amgorithm
(University of Eloued جامعة الوادي, 2022-01-24) Benyoucef, Aicha; Hamadouche, M’Hamed
In recent years, hiding information in medical images is the largest usage to secure this information or garnet the integrity of the owner. But this embedding can distort the medical image and change the necessary patient information. In this paper, we propose a robust method for medical image watermarking; first of all. The original image is filtered by a sharpening filter for enhanced contrast then separate their two regions using snake segmentation. The embedding mark (an electronic patient record) is added to the frequency domain after applying Discrete Wavelet Transform (DWT) on the region of non-interest (RONI) using the last signification bit (LSB). This region has a predominantly black background where a region of interest (ROI) has the necessary patient information. This method preserves a high-quality watermarked image, and it improves authentication. Our method is evaluated by Pick Signal to Noise Ratio (PSNR = 46.4039 for 512*512 bits image size) SNR, NC, and Histogram analysis.
ParPredict: A partially-ordered sequential rules based framework for mobility prediction
(University of Eloued جامعة الوادي, 2022-01-24) Amirat, Hanane
Predicting the future movement of mobile users has emerged as an important technology topic in many applications related to intelligent transportation systems (ITS) and Location-based services (LBS). Numerous prediction models were proposed relying on probabilistic models (e.g. Markov Chain) or data mining techniques (e.g. neural network, sequential patterns mining). Mining sequential patterns and rules is one of the data mining techniques used. Mining sequential rules from sequence databases is an active research topic that is broadly applied for many real-world scenarios. In this paper, we propose to adapt a novel kind of sequential rules called partially order sequential rules for route prediction problem. We aim to further compare this kind with standard sequential rule for the task of mobility prediction. An experimental evaluation conducted on real and synthetic datasets show that the proposed model outperforms a state-of-the-art sequential model in terms of accuracy and prediction coverage.
Ball bearing monitoring using decision-tree and adaptive neuro-fuzzy inference system
(University of Eloued جامعة الوادي, 2022-01-24) Euldji, Riadh; Boumahdi, Mouloud; Bachene, Mourad; Euldji, Rafik
This study aims to provide a methodology that relies on the combination of the following approaches: the decision tree, the neural network, and the fuzzy logic to monitor the evolution of bearing degradation. Data collected from the vibratory signals generated from the tests carried out on ball bearings mounted in an experimental fatigue platform, are used. The decision tree method is applied to select the most relevant monitoring indicator, which will be used to develop an Adaptive Neuro-Fuzzy Inference System (ANFIS). The training and test data required for model development have been classified according to the following states: normal, abnormal, and dangerous. These were defined from two thresholds: alert threshold and danger threshold. Then, the ANFIS model is trained from the indicators selected by the decision tree to predict the behaviour of the bearing in operation. The results confirm the effectiveness of the proposed approach for monitoring the health of ball bearing
Multi Agent Systems based CPPS – An Industry 4.0 Test Case
(University of Eloued جامعة الوادي, 2022-01-24) Bendjellou, Abdelhamid; Gaham, Mehdi; Bouzouia, Brahim; Moufid, Mansour; Mihoubi, Bachir
With the rise of the Industry 4.0 Revolution, Artificial Intelligence, digitalization, and connectivity have been more than ever; adopted in the industrial world. This adoption is leading to the transformation of the mechatronic systems used in production into Cyber-Physical Production Systems. Such a concept is taking industrial Automation and computer integrated manufacturing to the next level. The massive migration of traditional production systems into Cyber-Physical Production Systems, including the MAS-based CPPS, made the reviewing of the traditional methods of Engineering and Commissioning a must. Which explains the increase in the number of research works during recent years about the application of these Architectures on practical cases. In the present paper, we propose a way of developing and implementing MAS-based CPPS on an Industry 4.0 Assembly Platform. Moreover, we test the behavior of the Multi-Agent systems with interaction with SIEMENS Programmable Logic Controllers via OPC UA Protocol, during a Software-In-the-Loop “SIL” Test on a 3D Model of the Platform running on a separate Computer. The test assesses the behavior of the components of a typical Cyber-Physical production module during the treatment of a given operation on the product, to extract the vulnerabilities in the treatment of the operation and search for appropriate improvements.
Dilated Convolutions based 3D U-Net for Multi-Modal Brain Image Segmentation
(University of Eloued جامعة الوادي, 2022-01-24) Kemassi, Ouissam; Maamri, Oussama; Bouanane, Khadra; Kriker, Ouissal
Several deep learning based medical image segmentation methods use U-Net architecture and its variants as a baseline model. This is because U-Net has been successfully applied to many other tasks. it was noticed that the U-Net-based models are unable to extract features for segmenting small masks or fine edges. To overcome this issue, we propose a new 3D U-Net-based model, baptized Y- Net. In this model, we make use of dilated convolution which has shown its effectiveness in grasping different features at different scales. This allows us to capture more information from small anatomical parts. Our model is assessed on MRbrains13 dataset for brain tissue segmentation task. Compared to the traditional UNet 3D, the obtained results show that the proposed model performs well, especially in segmenting white matter and grey matter tissues.
Multi-agent system for an adaptive intrusion detection
(University of Eloued جامعة الوادي, 2022-01-24) Cheikh, Mohamed; Hacini, Salima
The denial-of-service attack is to make a service unusable; this can be done by overloading the network with useless information, generally leveled against application servers or web servers. The intrusion detection systems are powerful tools for the detection of attempted attacks DOS (Denial of Service). However, they suffer from a number of problems such as high rate of false positives and negatives. In this paper, we present a new mechanism for intrusion detection DOS, based on the use of adaptive agents. This self-learning mechanism guaranteed DOS attack detection with minimal false alarms.
Increasing the performance of mobile networks by planning and dimensioning
(University of Eloued جامعة الوادي, 2022-01-24) AYAD, Mouloud.; Medjedoub, Smail; Benziane, Mourad; Saoudi, Kamel; Arabi, Abderrazak
The telecommunications systems are very widely present in our recent smart lifestyles. The key element of these systems is to ensure the high quality of service. For this purpose, the operators integrated new services and technologies. To evaluate these services, the operators have adapted their planning methods. This will allow an ample range and exemplary service to keep the latent augmentation in traffic. For this reason, the need for a crucial and inevitable tool for increasing the performance of mobile networks is essential. The main objective of this work is the amelioration of the network's performance by planning, optimization and dimensioning techniques of a mobile network. To achieve this objective, this study is based on real measurements of the different data (Received Signal Received Quality (RSRQ), Reference Signal Received Power (RSRP), and Signal to Interference & Noise Ratio (SINR)).
A Smart Home Management based on M2M/IoT Technologies
(University of Eloued جامعة الوادي, 2022-01-24) Djehaiche, Rania; Aidel, Salih; Benhamimid, Karima
The advances in the applications based on Machine-to-machine (M2M) and Internet of Things (IoT) enabled the evolution of Smart home solutions wirelessly. This paper aims to present the implementation of smart homes based on M2M/IoT technologies using wireless sensor networks (WSNs). In this system, Arduino Uno microcontroller has been used with several compatible sensors, actuators, and modules to make the control and take the suitable decision. In addition, various technologies of communication such as Bluetooth, Wi-Fi, Ethernet, and GSM are used as wireless communication mediums to enable the interaction between users and the proposed system. Our Smart home implements integrated functions and services like security, indoor care, and outdoor care. A simple smartphone remotely controls these functions through our mobile application named ‘Raniso’. The experimental results of using the proposed system show that a variety of events can be detected and monitored efficiently. Gas leakage, fire, and housebreaking situations can be detected and users get notified about them via Calls and messages. Also, lights, temperature monitoring, and fans can be easily controlled remotely. Besides, the proposed system can perform some proper actions including decreasing gas concentration via the opening of windows, watering the garden when the soil is dry, saving the rainwater, and so on. Our smart home system is very utile to prevent losses in resources and human life caused by undesired events as also it conserves energy and provides comfort.
Applying Artificial intelligence techniques for predicting amount of CO2 emissions from calcined cement raw materials
(University of Eloued جامعة الوادي, 2022-01-24)
This paper aims to predict the amount of carbon dioxide CO2 emissions from raw material used in cement clinker production during calcination process. The amount of CO2 emissions is mainly from the decarbonisation thermal process that is directly related to chemical composition, distribution of particle size and time exposed at high temperature. These influencing factors interact with each other, making the calculation of the amount of CO2 emissions with conventional techniques more difficult. For this reason, several artificial intelligence techniques are applied to predict the amount of CO2 emissions. The key advantage of the proposed techniques is its ability to learn and to generalise without any prior knowledge of an explicit relationship between target and its influencing parameters. The intelligence techniques applied are deep neural network (DNN), artificial neural networks (ANN) optimised using ant colony optimization (ACO-ANN) and genetic algorithm (GA-ANN). The results obtained are promising and show that all intelligence techniques can provide excellent accuracy with high R2 and low error. DNN predicts the amount of CO2 emissions very accurately when comparing to other techniques. Overall, the performance accuracy of ACO-ANN technique is higher than the GA-ANN. According to R2 values, there are more than 99%, 98.5% and 98% of experimental data in testing phases can be explained by DNN, ACO-ANN and GA-ANN respectively with average relative error less than 1.04%. As conclusion, all intelligence techniques can be employed as an excellent alternative to predict the amount of CO2 emissions.
New Approach for Multi-Valued Mathematical Morphology Computation
(University of Eloued جامعة الوادي, 2022-01-24) L'haddad, Samir; Kemmouche, Akila
Mathematical Morphology (MM) is a useful tool for spatial image processing. It is based on an infimum operator (min) and a supremum operator (max) applied in local neighborhoods to detect pixel extremes. The MM was initially defined for mono-band images in which each pixel image is a scalar value and it is easy to find pixels extremes by the infimum and the supremum operators. However, in the case of multi-band images, where each pixel image is represented by a vector, establishing an order between image pixels in local neighborhoods by the infimum and supremum operators is not trivial. Many works discussed the feasibility to extend the MM to multi-band images but they did not lead to any consensual definition of the multi-valued mathematical morphology. Nevertheless, these existing works agree that the definition of the MM for multi-band images is based on the notion of vector ordering. In this paper, we propose a multi-valued MM operators computing by introducing a new vector ordering algorithm that allows extending the scalar MM to multi-band images. The proposed multi-valued morphological operations were tested in the experimental phase for the morphological descriptors computation. The obtained results based on use of the proposed vector ordering algorithm for the multi-valued MM computing improve the classification rates.
Assessing Combinations of Artificial Neural Networks Input Meteorological Parameters to Improve Daily Runoff Simulation
(University of Eloued جامعة الوادي, 2022-01-24) Aoulmi, Yamina; Marouf, Nadir; Amireche, Mohamed
Hydrological models are one of the tools used for reconstruction or simulation, forecasting for the anticipation of future changes in the flow of a river, which allows better management of water resources during low flow periods and anticipation of flood risks during high water periods. Using artificial neural network (ANN) in the fields of hydrology and water resources, became advocated due to its capability of tackling, modeling and forecasting the problems that are nonlinear or stochastic within the Rainfall-Runoff (R-R) system. This research aims to test the practicability of using ANNs with two input configurations; to model the R-R relationship in two stations Mirebek and Ain Berda located in the Seybouse basin in Algeria. The 1st (ANN IP) considers only precipitation as an input variable for the daily flow simulation. The 2nd (ANN IIP,T,Hu) considers a combination of the Temperature and Humidity with precipitation at the model input. The results of the two models were compared through performance metrics, viz., Root Mean Square Error (RMSE),mean absolute error (MAE), Pearson's correlation coefficient (R), Nash Sutcliffe Efficiency coefficient (NSE), and through graphical interpretation (scatter plots and time series). Better flow simulations were provided by the three-input model for the two stations; where R=0.90, NSE=80.5% for Mirebek station and R=0.85, NSE=72% for Ain Berda. This result has confirmed that as much input variables are numerous, as more the model of ANN is efficient. The finding of this study indicates that the developed ANN models could be considered as a powerful tool for predicting Runoff.
Using Artificial Intelligence for Microgrid Operation and Control in Presence of Sustainable Power Systems
(University of Eloued جامعة الوادي, 2022-01-24) KOUBA, Nour EL Yakine; SADOUDI, Slimane
Nowadays, and with the growth use of renewable energy sources (RESs), an intelligent control scheme is requisite to ensure a stable power generation system especially in isolated and small areas. In the classical grid, the load frequency control loop (LFC) is widely used to cope with sudden and random load variation. However, with the large share of RESs and due to their stochastic behavior, a robust controller is needed to suppress frequency deviation and handle system frequency at scheduled value. In this scope, this paper aims to propose the use of first time a recently population-based optimization algorithm named Sine Cosine Algorithm (SCA) as artificial intelligence strategy to design an optimal LFC controller that allow the integration of large amount of RESs in the microgrid (MG). A hybrid microgrid including diesel engine, wind farm and photovoltaic generator was investigated. The SCA algorithm was employed to optimize the PI controller parameters of the LFC loop. To promote the use of RESs and minimize the use of fuel, a Hybrid Energy Storage System (HESS) combined Redox Flow Batteries (RFB) and Superconducting Magnetic Energy Storage (SMES) was considered. The HESS was used to support the LFC loop and reduce frequency fluctuation in presence of RESs. To prove the validity of the proposed strategy, various scenarios have been simulated. In addition, a comparative study between the proposed SCA algorithm and some metaheurisctic algorithms such PSO and GA have been performed. Furthermore, robustness analysis have been carried out with different rate of RESs penetration. Finally, the obtained results demonstrate the effectiveness and superiority of the developed strategy for dynamic microgird control.
Big Data Veracity: Methods and challenges
(University of Eloued جامعة الوادي, 2022-01-24) benabderrahmane, MOUTASSEM; laouni, DJAFRI; GAAFOUR, Abdel-Kader
Today, the Internet has become the main source of information, a place where there are no restrictions on who can share information .This latter can play an important role in prediction, estimation and decision making processes. But, this role will not only be achieved through abundance, it will also be the result of data quality. Veracity refers to the assurance of quality or credibility of the data collected. The data can be incomplete, biased, vague or wrong. For this reason, automatic filtering mechanism has been developed. Moreover, due to the increasing velocity of information spread, manual assessment of information veracity became hard, a time consuming process and even the already existing automatic filtering mechanisms has to be improved to cope with the speed of information spread. In this paper, a literature review is established to highlight the recent methods and techniques which are exploited in computerized veracity assessment. The challenges and limitations of existing works will be discussed, and future research directions will be proposed to address critical issues of data veracity in the era of big data.
A hybrid LBP-HOG model and naive Bayes classifier for knee osteoarthritis detection: data from the osteoarthritis initiative
(University of Eloued جامعة الوادي, 2022-01-24) Messaoudene, Khadidja; Harrar, Khaled
Knee OsteoArthritis (KOA) is a disease characterized by a degeneration of cartilage and the underlying bone. It does not evolve uniformly; it can stay silent for a long time and can quickly intensify for several months or weeks. For this reason, it is necessary to develop an automatic system for diagnosis and reduce the subjectivity in the detection of the disease. In this paper, we present a method for detecting knee osteoarthritis based on the combination of histograms of oriented gradient (HOG) and local binary pattern (LBP). Four classifiers including KNN, SVM, Adaboost, and Naïve Bayes were tested and compared for the prediction of the illness. A total of 620 X-Ray images were analyzed, composed of 310 images from healthy subjects (Grade 0), and 310 images from pathological patients (Grade 2). The results obtained reveal that Naïve Bayes achieved the highest performance in terms of accuracy (ACC = 91%) on the Osteoarthritis Initiative (OAI) dataset. The fusion of HOG and LBP features in KOA classification outperforms the use of either feature alone and the existing methods in the literature.
Recognizing Arabic handwritten literal amount using Convolutional Neural Networks
(University of Eloued جامعة الوادي, 2022-01-24) KORICHI, Aicha; SLATNIA, Sihem; TAGOUGUI, Najiba; ZOUARI, Ramzi; KHERALLAH, Monji; AIADI, Oussama
Currently, deep learning techniques have become the core of recent research in pattern recognition domain and especially for the handwriting recognition field where the challenges for the Arabic language are stilling. Despite their high importance and performances, for the best of our acknowledge, deep learning techniques have not been investigated in the context of Arabic handwritten literal amount recognition. The main aim of this paper is to investigate the effect of several Convolutional Neural Networks CNNs based on the proposed architecture with regularization parameters for such context. To achieve this aim, the AHDB database was used where very promising results were obtained outperforming the previous works on this database.
Multi-Robot Visual Navigation Structure based on Lukas-Kanade Algorithm
(University of Eloued جامعة الوادي, 2022-01-24) ELASRI, Abdelfattah; CHERROUN, Lakhmissi; NADOUR, Mohamed
This paper presents an efficient control structure of two mobile robots based-visual navigation methods in an indoor environment. The proposed navigators are based on decision systems employed the necessary values estimated by a Lukas-Kanade (LK) algorithm of optical flow (OF) approach. The robots control systems use the generated motion values in order to detect and estimate the positions of the nearest obstacles and objects around each mobile robot. The multi-robot system task is to navigate autonomously in their environment safely without collisions. Obstacles are identified and detected with the employed cameras of each robot based on video acquisition and image processing steps. The efficiency of the proposed approach is verified in simulation using Visual Reality Toolbox. Simulation results demonstrate that the visual based control system allows autonomous navigation without any collision with obstacles..
A Novel Separable Convolution Neural Network for Human Activity Recognition
(University of Eloued جامعة الوادي, 2022-01-24) Boudjema, Ali; Titouna, Faiza
The issue with the time series classification arises in several human applications such as healthcare, industrial monitoring and cybersecurity. Recently, various methods have been developed in order to deal with this matter. In this paper, a novel deep learning-based model for human activity recognition is developed. The proposal examines deeply the training phase in which the acceleration metric is considered by exploring all components of the model. To this end, the architecture of the Convolutional Neural Network (CNN) is studied: a) first, we employ a separable CNN, where we integrate a particular filter model for the depthwise convolution; b) second, we combine the extracted features with the handcrafted features. The proposed classifier is evaluated using a human activity recognition dataset and compared to a set of recent works. The obtained results show that our model outperforms the compared methods under various metrics.

Browse

Browsing 03.Forum of Artificial Intelligence and Its Applications by Issue Date