International Journal of Computer (IJC - Global Society of Scientific Research and Researchers, GSSRR)
Not a member yet
459 research outputs found
Sort by
An Automated Method for Brain Tumor Segmentation Based on Level Set
Get PDF In this paper, an automatic method has been proposed for tumor segmentation. In this method, a new energy function by introducing the feature tumor is determined implemented by level set. Multi-scale Morphology Fuzzy filter is applied to the image and its output determines the tumor feature. The initial contour selection is important in active contour models. Therefor the initial contour has been selected automatically by using Hough transform and morphology function. Experimental results on MR images verify the desirable performance of the proposed model in comparison with other methods
The Least Significant Two-bit Substitution Algorithm for Image Steganography
Get PDFSteganography hides various data within different file types, which ensures secure communication. In recent years, the science of steganography has gained importance, due to the increase of large data on the internet and the safe transmission of these data. The main objective is to hide a large amount of data into the cover image in secure and incomprehensible manner. In this article, the two-bit least significant substitution (LSB) method is used to hide data on the colored cover images. In experimental studies, two-bit LSB substitution algorithm was performed in the form of LSB R2G2, LSB R2B2 and LSB G2B2 methods. In the classical LSB substitution method, the data are hidden in sequence, while the data are hidden by the shuffling algorithm at the proposed study. In this way, the security of hidden data is provided
WSN and Fuzzy Logic for Flash Flood and Traffic Congestion Detection
Get PDFFloods are the most common natural disaster and source of significant damage to life, agriculture and economy. Flash Floods are particularly deadly because of short timescales on which they occur. Most flood casualties are caused by a lack of information. There is no dedicated flood sensing systems that monitor propagation of flash floods in cities. .Human being do not have power to totally uproot natural calamity but they can predict natural calamity & take major steps to prevent it. Wireless Sensor Network (WSN) and Internet of Things (IoT) technology is used for predicting & detecting flooding condition in this study. WSN is preferred due to its cost effectiveness, faster transfer of data & accurate computation of required parameter for flood prediction. IoT combines embedded system hardware techniques along with data science or machine learning models. The model uses a mesh network connection over ZigBee for the WSN to collect data, and a GPRS module to send data to the internet. Data sets are evaluated using fuzzy logic to detect floods then broadcast alerts. Floods rarely occur hence the system is dedicated for traffic congestion notifications
Selecting the Honeywords from Existing User’s Passwords Using Improved Hashing and Salting Algorithm
Get PDFNowadays, hashing passwords become the most essential tool for various web applications for making login process. However, password hashing takes many times for processing and it has become easier for attackers to crack hashing passwords from legitimate users by using brute force attack. Brute force attack is one of the dangerous attacks for password hashing techniques. Therefore, the legitimate user accounts are stored the passwords with honeywords using honeywords generation algorithm in order to prevent from brute force attack. Honeywords generation method is to produce the fake or decoy password for deceiving the attackers. However, the existing honeywords generation algorithm meets the storage overhead problem. So, we are implementing the improved honeywords generation method which decreases the storage overhead problem and also it addresses the majority of the drawbacks of existing honeywords generation methods. Moreover, we store the password and honeywords into the database using a unique hashing algorithm with very low time complexity as most of the steps involved simple binary operations
Combination of Multiple Acoustic Models with Multi-scale Features for Myanmar Speech Recognition
Get PDFWe proposed an approach to build a robust automatic speech recognizer using deep convolutional neural networks (CNNs). Deep CNNs have achieved a great success in acoustic modelling for automatic speech recognition due to its ability of reducing spectral variations and modelling spectral correlations in the input features. In most of the acoustic modelling using CNN, a fixed windowed feature patch corresponding to a target label (e.g., senone or phone) was used as input to the CNN. Considering different target labels may correspond to different time scales, multiple acoustic models were trained with different acoustic feature scales. Due to auxiliary information learned from different temporal scales could help in classification, multi-CNN acoustic models were combined based on a Recognizer Output Voting Error Reduction (ROVER) algorithm for final speech recognition experiments. The experiments were conducted on a Myanmar large vocabulary continuous speech recognition (LVCSR) task. Our results showed that integration of temporal multi-scale features in model training achieved a 4.32% relative word error rate (WER) reduction over the best individual system on one temporal scale feature
Reducing Complexity of Java Source Codes in Structural Testing by Using Program Slicing
Get PDFStructural testing is one of the techniques of software testing. It tests only the structure of the source code while comparing expected results and actual results. Generally, structural testing takes a long time to perform its task and not possible. Sometimes, only a small portion of the program is relevant. This can be done by program slicing. Program Slicing is to decompose the program into smaller units that depends on different types of dependencies between the program statements. The different types of program slicing are forward slicing, backward slicing, complete slicing, dynamic and static slicing, etc. Moreover, there is Tree Slicing which is also a key technique to slice and merge different Symbolic Execution (SE) sub-trees under some specific conditions. In this paper, we combine Tree Slicing technique and Indus Kaveri where Indus is a robust framework for analyzing and slicing concurrent Java programs, and Kaveri is a feature-rich Eclipse-based GUI front end for Indus slicing. Then we present the experimental results in order to reduce the complexity of the java source code
Holistic Approach to Big Data Definition using Analysis of Facts
Get PDFBig data has become a concern of science, industry, business, and academics, thus it is no more a buzzword but an emerging technology as viewed by researchers from different perspectives. Thus, different perspectives produced different definitions, of which none of them fully described big data. This research analysed some profound definitions based on discovered facts about big data. The facts are its characteristics, its technology, mode of transfer, its analysis, its infrastructure and security. Thus, a new definition was proposed which captures the basic facts about big data. Big data has its characteristics as foundations, and the rest facts as the pillars. These facts reflect in-depth meaning and understanding of big data to science, industry, business and academics
Evaluating the Effectiveness of Digital Inclusion at Private Educational Schools in Gaza Strip
Get PDFThe research aims to evaluate the process of digital inclusion and the usage of digital tools in the educational process at private schools in Gaza Strip - Palestine. Education is the beacon of knowledge and development that is accompanied by developmental processes aimed at rising and improving the quality of service through the use of digital tools in all aspects of the educational and administrative processes within a school. In this study, the research team sought to know the tools used in the process of digital integration, and the challenges and problems faced by teachers while using digital tools, in addition to providing recommendations to help solve or mitigate problems. The descriptive analytical method is adopted to carry out the research. The results of the study showed weakness in the effectiveness of the digital inclusion process and the use of digital tools in the educational process in private schools. Furthermore, recommendations to the schools’ administrations, teachers and the Ministry of Education are stated in order to alleviate the problems and challenges of using digital tools
An Analysis of the Business Accelerator Programs in Turkey
Get PDFThis paper analyses the business accelerator programs in Turkey. Business accelerators are new generations of incubation programs born especially to support technology entrepreneurs and help them reach to the next level. There are six startup accelerators in Turkey out of eighteen which fit into the criteria of accelerator programs. These are Kworks, ITU Seed, SuCool, IOT Telco Labs, Pilot and Starter’s Hub programs. All of these programs only accept technology entrepreneurs and help them grow their businesses. Using the interview method, this study provides an inside look into the models of these six programs. It provides detailed analysis about the general structure of the programs, the characteristics of the entrepreneurs in the programs, how the programs operate, information about the graduates of the programs, the mentor networks, the investment possibilities and the performance criteria of the programs. There are studies about accelerators in highly developed countries but the literature lacks information about accelerators in developing countries. Therefore, this study contributes to the literature by filling this gap
Predicting Students\u27 Degree Completion Using Decision Trees
Get PDFEducational Data Mining (EDM) helped institutions to improve students\u27 performance by predicting student\u27s future learning behavior. To benefit from this, the researchers conducted this study to predict the successful degree completion and provide early intervention as necessary. Decision Tree algorithm provided by WEKA is used to build the model using students\u27 data such as Entrance Exam Results, gender, school type where they graduated high school and final grades from English 1, Algebra and major subjects. Students who entered the University from school years 2012-2013, 2013-2014, 2014-2015 and 2015-2016 were selected. RandomForest suited best for the model and desktop application was designed and evaluated as Outstanding in terms of Efficiency, Accuracy and User Friendliness.