The importance of the image analysis with respect to industrial, medical, satellite image processing applications is gaining attention of many researchers in recent times. The recognition of faults present in the damaged images is vital for based applications. In this paper, we aim at developing a method for identifying faults that present in images. Our approach is based on the concept of Bit Plane Filter using convex hull methods. The bit plane filtering methods used to slice the given images to fix on the affected portion of the given images. The convex hull method is used to identify the control points that are needed for reconstruction of images. The performance of bit plane method is evaluated using simulation and it is proved that our approach produces better results when compared to current methods.

Bit Plane Filter, Convex Hull, 3D Images, Medical Images.

The objective of the given work is to study the queues resulting in the buffer while self-similar traffic passing through a network nod and to build a mathematical model of an actual traffic. The latest researches of different types of network traffics bring out clearly that network traffic is defined by self-similarity and long-term dependence. The self-similar traffic has specific structure being reserved at various measures – its realization is characterized by some vast emissions when respecting low medium-scale traffic. This fact degrades the performance significantly (increases, losses and delays) while running through the network nods. Hence, it follows that commonly used methods of simulation and network system calculations rested on traditional assumptions do not reflect the real situation taking place in the network.

self-similar, network traffic, stochastic processes, queue.

A buffer is a region of physical memory storage that holds a specific amount of data and when that capacity is reached, the data will overflow. A buffer overflow is a condition where in the data transferred to a buffer exceeds the storage capacity. Buffer overflow is susceptible to attacks therefore, creates a necessity for intrusion detection systems. This paper combines code pointer integrity with buffer overflow visualization to address overflows attacks. During monitoring and detection stage, users are requested to introduce some data at a time and when an overflow is detected, the system will process and display appropriate buffer status. Colors are used to represent the different level of operations. A model to analyze the buffer behavior of the random process is derived. The effectiveness of our technique is confirmed through a visualization experiment. Result shows that transmission process is suspended when buffer overflows thus facilitate intrusion analysis.

Buffer-overflow, visualization, model and integrity check.

In this study, we introduced a novel feature ranking technique applicable to two well known classifiers Bayesian Belief network and Random Forest as both of these classification systems have been shown to be sensitive to the initial ordering of the features. We have illustrated that improvement in classification can be obtained even without ceding variables for feature (attribute) ranking sensitive classifiers. We also performed a comparison between Bayesian Belief network and Random Forest classification approaches in the well known feature subset selection and feature ranking problem. The proposed technique Polarization Measure (herein known as PM) is originated from within joint probability to discover the degree of explanation made by first feature (attribute)’s state to explain the other feature’s state. The technique has significantly better well performed in Bayesian belief network and better in random forest classifier in comparison to five feature ranking techniques and three well established feature subset selection techniques.

random forests algorithm; machine learning; Bayes structure learning; ranked features.

Les Objets Pédagogiques OPs sont les constituants pédagogiques les plus importants dans une formation elearning. Leur indexation d’une manière efficace les rendra réutilisables et par la suite rentables. Plusieurs recherches ont été menées pour offrir une meilleure indexation selon des techniques très variées partant de l’indexation à base du nom de l’OP et arrivant à l’indexation à base du type du document (texte, son, image, …etc.) en passant par la catégorisation thématique ou même la granularité des OP. Rares sont les travaux qui ont abordé l’indexation selon un critère que nous estimons très important qui est l’aspect pédagogique de l’apprenant et ses préférences d’apprentissage. En effet, les OP peuvent être pédagogiquement classés et indexés pour répondre d’une manière efficace aux besoins spéciaux pédagogiques d’un apprenant en termes de compétences à développer. Dans cet article, nous décrivons le mode d’indexation des compétences à développer par un OP dans le système de scénarisation et d’apprentissage à distance eMouss@ide [Ben11]. Ce système permet aux enseignants d’indexer les compétences à développer par un OP en utilisant le standard LOM - fr et par la suite de les rechercher pour une scénarisation d’un parcours de formation. Pour adapter le standard LOM-fr à ces exigences nous proposons de le compléter par d’autres sous catégories.

E-Learning, Compétences cognetives, indexation, Objets pédagogiques

Heart disease is the leading cause of death in India and worldwide. India is in the middle of a major economic and industrial transition. The life style changes have led to rise in hypertension, obesity, smoking, diabetes and in turn heart disease. Disease diagnosis often done based on doctors experience and personal opinion rather than the data hidden in the medical data base, which leads to wrong diagnosis and increases diagnosis costs which in turn affects the quality of services provided by hospitals to the patients. Medical data mining is to search knowledgeable data for effective medical diagnosis.K nearest neighbor is one of the widely used data mining technique in classification. It is a straight forward classifier where samples are classified based on the class of their nearest neighbor .Medical data bases are high volume in nature. If medical data contains redundant, irrelevant attributes classification will produce less accurate results. Feature subset selection is a dimensionality reduction technique used to remove redundant features and to increase accuracy. By applying feature subset selection on medical data we can determine the attributes which contributes more towards the disease which indirectly reduces no. of clinical tests to be taken by a patient .This paper investigates to apply K nearest neighbor with feature subset selection in the diagnosis of heart diasease. The experimental results show that applying feature subset selection to KNN will enhance the accuracy in the diagnosis of heart disease for Andhra Pradesh population.

Andhra Pradesh, Data mining, feature subset selection, heart disease, k-nearest neighbor, symmetrical uncertainty of attributes.

Texture plays an important role in numerous computer vision applications. Many methods for describing and analyzing of textured surfaces have been proposed. Variations in the appearance of texture caused by changing illumination and imaging conditions, for example, set high requirements on different analysis methods. In this thesis, methods for extracting texture features and recognizing texture categories using grey level first-order and second-order statistics, edge detectors and local binary pattern features are proposed. Unsupervised clustering methods are used for building a labeled training set for a classifier and for studying the performances of these features.Texture plays an important role in numerous computer vision applications. Many methods for describing and analyzing of textured surfaces have been proposed. Variations in the appearance of texture caused by changing illumination and imaging conditions, for example, set high requirements on different analysis methods. In this thesis, methods for extracting texture features and recognizing texture categories using grey level first-order and second-order statistics, edge detectors and local binary pattern features are proposed. Unsupervised clustering methods are used for building a labeled training set for a classifier and for studying the performances of these features.

texture analysis; classification; grey level statistics; local binary pattern.

Wireless capsule endoscopy is the golden standard in the investigation of the small bowel, but is time consuming and burdening for the physicians, as the imaging conditions are challenging and the variability of the acquired frames is great. This paper carries out a study on the automatic detection of bubbles and debris, based on an algorithm combining colour slicing techniques, textures and position within the WCE frame and the intestinal tract. Their identification represents an important phase in the investigation of a video file furnished by the endoscopic capsule, as it allows the reduction of the important frames submitted to analysis for potential lesion detection. Experimental results prove that the algorithm is effective in reducing the number of WCE images with relevant content, reducing thus the time spent for the analysis of the images acquired by WCE.Wireless capsule endoscopy is the golden standard in the investigation of the small bowel, but is time consuming and burdening for the physicians, as the imaging conditions are challenging and the variability of the acquired frames is great. This paper carries out a study on the automatic detection of bubbles and debris, based on an algorithm combining colour slicing techniques, textures and position within the WCE frame and the intestinal tract. Their identification represents an important phase in the investigation of a video file furnished by the endoscopic capsule, as it allows the reduction of the important frames submitted to analysis for potential lesion detection. Experimental results prove that the algorithm is effective in reducing the number of WCE images with relevant content, reducing thus the time spent for the analysis of the images acquired by WCE.

Wireless Capsule Endoscopy, medical image analysis, image segmentation, artefact detection, artificial neural network.

Confirmation of clinical breast cancer is done histopathologically on microscopic slides taking into consideration cell modifications, architecture modifications and mitotic cell index. This mitotic cells count (multiplying cells) is a separate index represented by counting the number of multiplying cells figures in at least 10 high power magnification microscopic fields (20x, 40x). It is a time consuming and demanding method, with low reproducibility, so it is suitable for an automated or semiautmated (computer aided) method. We have developed a method for mitosis figure identification using a detector based on a k-Nearest Neighbor algorithm applied on apixel-based feature detection. The results came with high sensibility, but with low specificity.

mitosis detection, computer aided diagnosis, semi automated mitotic index.

The convergent or convergent – divergent nozzle are often used in gas-dynamic technique. The study of these nozzles is complex and it requires complex mathematical models. The paper generates a Laval nozzle contour model 1 for critical diameter of 20 mm, that are known angle of convergence q1, divergence angle q2 and end point abscissa. Using gas dynamic functions determine ideal fluid motion parameters Laval nozzle in the convergence and divergence. The stagnation pressure is P0=6x10^6 Pa and the stagnation temperature T0=600 K. Supersonic air flow inside Laval nozzle is model using Fluent software.

Fluid Mechanics, Dynamics of gas, Laval nozzle.

We show the power of the simultaneous usage of GeoGebra and Maple for generalizing and proving of geometry problems. We present a simple school problem, where with the help of the dynamics in GeoGebra new geometric properties are recognize and then we prove them with the help of Maple. We state an open problem for an investigation. We suggest a new construction for GeoGebra that can optimize the construction process in the extended Euclidian plane.

Dynamic Geometry Software, Computer Algebra Systems, reflection about point, conic section, loci, homogenous coordinates.

The Monte Carlo method is being increasingly used in various fields like mathematical finance, engineering, physical sciences, and bioinformatics to solve problems where using the deterministic method is infeasible. The Monte Carlo simulation is a numerical computational technique which uses thousands or even millions of random values to solve complex problems, causing this technique to be slow and computer-intensive when being used for options pricing. Therefore, financial firms are usually forced to deploy powerful hard-ware means such as supercomputers and computer clusters in order to perform the needed simulations. The ability to link the traditional computers and servers available at organizations to form a grid that acts as a supercomputer can enable both scientists and professionals to run their simulations without spending extra costs. This paper proposes and implements an application for the Monte Carlo methods for options pricing using the enterprise grids on the Windows environment. The paper also provides a comparison between the performance of the proposed framework and the traditional model.

Financial Analysis, Monte Carlo Simulation, Option Pricing, Grid Computing, Enterprise Grids, Alchemi, .NET Framework.

The development of soft-computing approaches in law enforcement agencies has a wide range of application to intelligence analysis in the face of interrogation and evidence gathering during investigation. The capability of these tools to alleviate attempted deception by an informant or suspect is affected by many issues including the type of critical method, the type of hybridization used, and the ability to address issues of source reliability and information credibility. This work at the end will presents a hybrid model using neuro-fuzzy system to aid in the detection of deception during interrogation.

Interrogation, Investigation, Deception, Informant, suspect.

The rapid development of computer network especially the internet has enabled new software capabilities and wide market interest. The reality that this technology will virtually place all computer users on the internet is a major motivating factor in this work. This paper proposes a software piracy prevention technique in which client/user purchases a software product, but must connect to the remote server of the software developer before installation. The user provides an activation code that activates mobile agent. The validity of the activation is checked, the software user identity information is compared with store information in the database of the developer to ascertain authenticity and prevent piracy. A mathematical model is adopted for measuring the effectiveness of mobile agent migration in terms of network load which is expected to be at minimal.

Software Piracy Prevention, Remote server of the software developer, software users’ identity information, Mobile agent.

Management issues in a network communications are complex, difficult and the size of the networks make network management a daunting task. It is a broad term that includes numerous complex mechanisms for optimizing network performance. Hence, a suitable and effective approach is needed in real time for resource utilization. This paper describes a novel idea of using Fuzzy Logic (FL) for effective network management. FL statistics is based on the likelihood function of the model which is applied to detect point of traffic failure by comparing the null hypothesis against the alternative hypothesis. In our approach, the structure of Fuzzy decision has two dimensions: Input/output. The inputs are the Hurst parameters and its changing rate, while the output is the influence of traffic intensity on network performance. We employed Entropy Measure to validate the functionality of our approach and the result demonstrates that the approach is sufficient for traffic classifications.

networks, management, self – similarity, Fuzzy Logic, degradation.

Providing network with sufficient bandwidth is necessary, but not sufficient, step to ensuring good performance of a network application. If excessive network Latency is causing the application to spend a large amount of time waiting for responses, then the bandwidth may not be fully utilized and performance will suffer. Thus, an important aspect of data communication is to estimate the rotational latency cause by I/O subsystem. In this paper, we present a single model structure that can be used to represent complex I/O subsystems at varying levels. We developed some parameters for estimating the contention components of the effective service demands under a number of different assumptions and incorporating these parameters into network modeling. The result demonstrates real world effects of latency using the time to load a web page. This clearly shows that latency directly affects the way a user obtains data from the internet.

latency, contention, Rotation, Disks, Channel, Control Unit.

This paper presents data on hydrocarbon combustion stoichiometry and algorithms for determining flame temperature. If that does not take into account the dissociative reaction products (CO2, H2O and N2) algorithm allows the calculation of the flame temperature of the heat balance equation. It analyzes the general case of dissociative-products CO2 and H2O. In this case we apply a different algorithm that solve a system of nonlinear equations with six unknowns vCO2d, vCOd, vH2Od, vH2d, vOHd, vO2d then apply the first algorithm to determine the flame temperature. When burning natural gas with the initial conditions t0 = 15 C and pressure of the atmosphere resulting tf = 2049.246 C under the same conditions, with consideration of dissociation, tf = 2049.822 C. Algorithms do not use iterative solution of the equation of thermal balance.

Stoichiometric combustion, Thermo-dynamics.

Today’s banking system are engulf in greater risks arising from the global economy recession. These banks are struggling to emerge from the economic recession and depression by capturing and retaining much more loyal and stable customers in the financial stage. This research work models and implement a Customer-based Knowledge Management System that will exploit customers’ knowledge for competitive advantage. This study utilized a Collaborative Customer Intelligence Architecture (CCIA) with a multidimensional approach to pragmatic product developments processes which permits data to be merged, transformed, analyzed and determine customer’s patterns or behaviour. The system utilized the Apriori algorithm of associative data mining to promote customer segmentation while recommending the most fitted products and services to individual customers. This system is called the Collaborative Customer Management System (CCMS). This paper addressed significant gaps in existing customer relationship landscape within the banking operations. The results showed how an intelligent product recommender would help the banks to roll out basic products and services that would transform customers from ordinary passive recipients of products and services to an empowered knowledge collaborators.

Customer Knowledge Management, banking industry, Apriori Algorithm, data mining.