International Journal of Computer and Information Technology
Not a member yet
138 research outputs found
Sort by
A New Computer Vision Based Rail Detection Method Using Entropy and Support Vector Machines
Get PDF Condition monitoring in railways is an important and critical process in terms of travel safety. However, this process is generally done based on observation or with various equipment. Therefore, it is costly and has a high probability of error. In this study, a computer vision-based method for rail detection for condition monitoring in railways is proposed. In addition to the features obtained from the images, a new feature is calculated using entropy. Rail detection is provided by classifying these features with Support Vector Machine (SVM). It has been seen that the proposed method works successfully and provides improvement in the monitoring process
Fraud Detection in Motor Insurance Claims Using Supervised Learning Techniques: A Review
Get PDFFraudulent claims have been a big drawback in motor insurance despite the insurance industry having vast amounts of motor claims data. Analyzing this data can lead to a more efficient way of detecting reported fraudulent claims. The challenge is how to extract insightful information and knowledge from this data and use it to model a fraud detection system. Due to constant evolution and dynamic nature of fraudsters, some approaches utilized by insurance firms, such as impromptu audits, whistle-blowing, staff rotation have become infeasible. Machine learning techniques can aid in fraud detection by training a prediction model using historical data. The performance of the models is affected by class imbalance and the determination of the most relevant features that might lead to fraud detection from data. In this paper we examine various fraud detection techniques and compare their performance efficiency. We then give a summary of techniques’ strengths and weaknesses in identifying claims as either fraudulent or non-fraudulent, and finally propose a fraud detection framework of an ensemble model that is trained on dataset balanced using SMOTE and with relevant features only. This proposed approach would improve performance and reduce false positives
Empirical Research on Realizing, Evaluating, and Validating a Conceptual Breast Cancer e-portal Model with Arabic Content
Get PDFThis paper presents an empirical research that realizes a previous research of a conceptual breast-cancer e-portal model with Arabic content. The paper starts with briefing the previous research, after that, it highlights the gaps need to be bridged and the problems need to be solved. A real development of an e-portal prototype is done for achieving the research goal. This e-portal prototype has applied the conceptual model of the previous research
Sorting Real Numbers in Constant Time Using n^2/log^cn Processors
Get PDFWe study the sorting of real numbers into a linked list on the PRAM (Parallel Random-Access Machine) model. We show that n real numbers can be sorted into a linked list in constant time using n2 processors. Previously n numbers can be sorted into a linked list using n2 processors in O(loglogn) time. We also study the time processor trade-off for sorting real numbers into a linked list on the PRAM (Parallel Random Access Machine) model. We show that n real numbers can be sorted into a linked list with n2/t processors in O(logt) time. Previously n real numbers can be sorted into a linked list using n3 processors in constant time and n2 processors in O(loglogn). And then we show that input array of n real numbers can be sorted into linked list in constant time using n2/logcn processors for any positive constant c. We believe that further reduction on the number of processors for sorting real numbers in constant time will be very difficult if not impossible
A Review of Smishing Attaks Mitigation Strategies
Get PDFMobile Smishing crime has continued to escalate globally due to technology enhancements and people\u27s growing dependence on smartphones and other technologies. SMS facilitates the distribution of crucial information that is principally important for non-digital savvy users who are typically underprivileged. Smishing, often known as SMS phishing, entails transmitting deceptive text messages to lure someone into revealing individual information or installing malware. The number of incidences of smishing has increased tremendously as the internet and cellphones have spread to even the most remote regions of the globe
The Evolution of Information and Communication Technologies: Towards uses oriented collaborative practices: Towards uses oriented collaborative practices
No full textIn all sectors of activity, knowledge is the most important strategic lever in the organizations’ management. The informational practices induced by Web technologies and data exploitation are accompanied by a paradigm that disrupts the process of sharing and communicating information in organization. Knowledge generated by human and social activities broadens the scope of knowledge management systems. New forms of technologies are changing system’s design approaches. It is on these evolving aspects of Information and Communication Technologies (ICT) that we will highlight the different challenges faced by organizations. In reviewing the literature on ICTs, several researches work highlight the importance of digital revolutions in the development of IS in general, and in particular Knowledge Management (KM) systems. In order to understand the evolution of KM Systems, we carry out the postulate that different angles of view have a positive correlation with our research perimeter: We identify that collaborative practices and decision-making approach are positively related to Knowledge Management systems’ design. The conceptual model links our hypothesis with the concept of knowledge management systems. We develop a conceptual model according to a review of literature in information science. Research finding can be used for designing Knowledge Management systems. Several actors can benefit from the repercussions on a pragmatic level. System designers can identify modern developments to provide more pragmatic applications. Decision-makers can identify how to harness Knowledge Management systems and also identify good practices in terms of collaborative approaches and decision-making processes
Combining Image Processing Techniques and Mobile Sensor Information for Marker-less Augmented Reality Based Reconstruction
Get PDFMarker-less Augmented Reality(AR) based recon- struction using mobile devices, is a near impossible task. When considering vision based tracking approaches, it is due to the lack of processing power in mobile devices and when considering mobile sensor based tracking approaches, it is due to the lack of accuracy in mobile Global Positioning System(GPS).
In order to address this problem this research presents a novel approach which combines image processing techniques and mobile sensor information which can be used to perform precise position localization in order to perform augmented reality based reconstruction using mobile devices. The core of this proposed methodology is tightly bound with the image processing technique which is used to identify the object scale in a given image, which is taken from the user’s mobile device. Use of mobile sensor information was to classify the most optimal locations for a given particular user location.
This proposed methodology has been evaluated against the results obtained using 10cm accurate Real-Time Kinematic(RTK) device and against the results obtained using only the Assisted Global Positioning System(A-GPS) chips in mobile devices. Though this proposed methodology require more processing time than A-GPS chips, the accuracy level of this proposed methodology outperforms that of A-GPS chips and the results of the experiments carried out further convince that this proposed methodology facilitates improving the accuracy of position local- ization for augmented reality based reconstruction using mobile devices under certain limitations
Classification of Facial Expression Using Principal Component Analysis (PCA) Method and Support Vector Machine (SVM)
Get PDFClassification is a process to assert an object into one of defined categories. This study examines the classification of recognition of student’s facial expression during digital learning –indifferent and serious expression. The dataset used was from a vocational school -SMK Muhammadiyah 2 Bantul. This study used the combination of algorithm: Principal Component Analysis (PCA) and Support Vector Machine (SVM) to increase the accuracy. This study aims at comparing the performance of combination of two algorithm: (PCA to SVM) and (PCA to k-NN). The result states that the combination of PCA-SVM algorithm is higher than the combination of PCA-k-NN algorithm with the average accuracy of 96% and 89%
A Framework for Verification in Contactless Secure Physical Access Control and Authentication Systems
Get PDFBiometrics is one of the very popular techniques in user identification for accessing institutions and logging into attendance systems. Currently, some of the existing biometric techniques such as the use of fingerprints are unpopular due to COVID-19 challenges. This paper identifies the components of a framework for secure contactless access authentication. The researcher selected 50 journals from Google scholar which were used to analyze the various components used in a secure contactless access authentication framework. The methodology used for research was based on the scientific approach of research methodology that mainly includes data collection from the 50 selected journals, analysis of the data and assessment of results. The following components were identified: database, sensor camera, feature extraction methods, matching and decision algorithm. Out of the considered journals the most used is CASIA database at 40%, CCD Sensor camera with 56%, Gabor feature extraction method at 44%, Hamming distance for matching at 100% and PCA at 100% was used for decision making. These findings will assist the researcher in providing a guide on the best suitable components. Various researchers have proposed an improvement in the current security systems due to integrity and security problems
Feature Extraction using Histogram of Oriented Gradients for Image Classification in Maize Leaf Diseases
Get PDFThe paper presents feature extraction methods and classification algorithms used to classify maize leaf disease images. From maize disease images, features are extracted and passed to the machine learning classification algorithm to identify the possible disease based on the features detected using the feature extraction method. The maize disease images used include images of common rust, leaf spot, and northern leaf blight and healthy images. An evaluation was done for the feature extraction method to see which feature extraction method performs best with image classification algorithms. Based on the evaluation, the outcomes revealed Histogram of Oriented Gradients performed best with classifiers compared to KAZE and Oriented FAST and rotated BRIEF. The random forest classifier emerged the best in terms of image classification, based on four performance metrics which are accuracy, precision, recall, and F1-score. The experimental outcome indicated that the random forest had 0.74 accuracy, 0.77 precision, 0.77 recall, and 0.75 F1-score