International Journal of Computer and Information Technology
Not a member yet
138 research outputs found
Sort by
Development of a Self-Organizing Multipurpose Mobile Robot
Get PDFDevelopment of a Self-Organizing Multipurpose Mobile Robot is aimed at creating a versatile robotic system capable of autonomously navigating through environments, monitoring surroundings, and streaming real-time data. The project focused on integrating advanced sensors and technology to achieve seamless obstacle avoidance, environmental data collection, and remote monitoring. It involved identifying the project requirements, selecting appropriate hardware and software components, assembling the chassis, mounting sensors, and coding for integration. The core hardware components included the ESP32-CAM module, flame sensor, PIR motion sensor, light detector sensor, rain sensor, and others. These components synergized to enable the robot to interact with its environment and make informed decisions. It further delved into the construction process, detailing the steps of assembling the chassis, mounting components, wiring sensors, and programming the microcontroller. The integration of various sensors was meticulously orchestrated to ensure accurate data collection and real-time video streaming. Subsequent testing and result analysis validated the project\u27s success in achieving its goals. The robot demonstrated efficient obstacle avoidance, precise environmental data acquisition, and stable remote video streaming. The project\u27s coding and system design facilitated seamless collaboration between hardware and software components, culminating in a functional and agile robotic system. The discussion and performance analysis section elaborated on the significance of sensor integration, coding efficiency, and seamless communication. In conclusion, the project\u27s success underscored the potential of mobile robotics in diverse applications. The project\u27s outcomes lay the groundwork for future innovations in robotic systems, emphasizing the importance of strategic integration, meticulous coding, and comprehensive testing in achieving functional and adaptable robots
Deep Learning Model for Crop Diseases and Pest Classification
Get PDFThe deep learning model for crop diseases and pest classification research examined how deep learning might improve farming methods, particularly for the purpose of accurately classifying pests and illnesses that affect crops. The importance of crop diseases and pests to world food security was highlighted in the introduction, along with the need for new approaches, such as deep learning models, to improve the accuracy and effectiveness of pest and disease control in farming. In order to evaluate the classification accuracy, the secondary datasets obtained from kaggle website were used to train and test various deep learning models, one of which being DenseNet. The researcher used a thorough assessment methodology to compare DenseNet\u27s performance to that of other models, including AlexNet, EfficientNet,Visual Geometry Group, and Convolution Neural Network.With an impressive accuracy score of 96.988% on the maize disease dataset and 96.9382% on the pest dataset, DenseNet proved to be the best model among the others. More accurate predictions were the result of DenseNet\u27s capacity to effectively collect intricate characteristics and patterns within the visual data, which led to its improved performance. The researcher examined the implications of DenseNet\u27s high accuracy in the discussion section, implying that its sophisticated design rendered it optimal for the categorization of agricultural diseases and pests. In addition, the researcher investigated the feasibility of incorporating DenseNet into practical agricultural systems, where its strong performance might greatly enhance methods of crop monitoring and disease control. The discussion came to a close with suggestions for future studies, such as looking at whether DenseNet can be used for other types of crops and if hybrid models or transfer learning may improve its performance
A Systematic Review of Predictive Factors for Learner Attrition in Online Learning: Insights for Machine Learning Models
Get PDFOver the past ten years, online education has expanded rapidly due to its accessibility, scalability, and flexibility. Despite its potential, high attrition rates in online education threaten both student progress and the legitimacy of the institution. A comprehensive analysis of empirical research on the factors influencing learner attrition in online learning settings is presented in this study. To identify the individual, course-level, institutional, and technical causes of attrition, it incorporates and categories the body of existing work. The results point to the complex aetiology of attrition and identify important domains for focused intervention and predictive modelling
Development of an NFC-Based Tracking System for Production, Inventory, and Maintenance in Manufacturing
Get PDFThis paper presents the development, implementation, and evaluation of a computational system for integrating production control, inventory management, and maintenance operations in a manufacturing company using Near Field Communication (NFC) technology. The proposed solution leverages NFC tags as a core element to ensure real-time tracking of products and components throughout the manufacturing process. The system architecture consists of NFC tags for product identification, a custom mobile application for data acquisition, and a centralized database for storage and analysis. The system workflow enables operators to register, monitor, and update the status of products using mobile devices, enhancing traceability, efficiency, and data accuracy. The solution was validated in a real-world manufacturing environment through functional tests and a System Usability Scale (SUS) questionnaire, which yielded an average usability score of 72.08, indicating a good level of user satisfaction. These results demonstrate the feasibility, usability, and practical benefits of the proposed system, which contributes to the digital transformation of industrial operations and aligns with Industry 4.0 principles
A Comparative Analysis of Deep Learning Models for Detection of Lumpy Skin Disease with emphasis on Shifted Window Transformers
Get PDFLumpy Skin Disease (LSD) in cattle is an increasingly prevalent viral infection with significant economic impact. Traditional detection methods are often labor-intensive and delayed. In this study, five state-of-the-art deep learning (DL) architectures—ResNet50, EfficientNetB0, MobileNetV2, Vision Transformer (ViT-B16), and Swin Transformer Tiny (Swin-T)—were evaluated and compared for image-based LSD classification. Publicly available Kaggle datasets of infected and healthy cattle were used. All models were fine-tuned using transfer learning and tested for classification accuracy, F1-score, inference time, explainability (via Grad-CAM), and real-world deployability. Results show that Swin-T achieved the highest classification accuracy of 95.3%, while MobileNetV2 emerged as the most deployment-friendly model. Grad-CAM visualizations confirmed that transformer-based models captured relevant lesion features with greater spatial sensitivity than CNNs. The study highlights the promise of hybrid transformer-CNN models for practical livestock diagnostics, especially in resource-constrained environments
Comparison with Deep Learning Methods For Predicting Stock Prices
Get PDFRecently, machine learning has been an essential tool for analysis in diverse fields, including science, sports management, and economics. In particular, the stock market comprises a complex network of buyers and sellers engaged in stock trading. So, predicting stock prices has been developed using machine-learning techniques to significantly enhance such forecasts\u27 accuracy. Recent advancements have improved the performance of several algorithms, such as Linear Regression, Support Vector Machines (SVM), and K-nearest neighbors (KNN) to predict stock prices. Stock price datasets typically contain information such as opening and closing prices, high and low values, dates, trading volume, and adjusted closing prices provided by Yahoo Finance. Based on the data, this research evaluates the prediction accuracy of each machine-learning method and presents the results through data visualizations, including box plots and tables. The compiled results will assist in identifying the most effective model for stock price prediction
A Novel Method for Secure Key-Sharing in 5G and Beyond
Get PDFIn today’s world, online communication is essential, and infrastructure for cutting-edge mobile technologies like 5G, is growing daily to meet the demand. So, information sharing security needs to be safeguarded as electronic communications spread. To implement this, cryptography is typically used, and most commonly symmetric key cryptography, due to its many advantages over other crypto-systems. However, one significant disadvantage of symmetric key system is that the single-key-sharing is exposed to all entities in a network communication system, which makes the subsequent communications vulnerable to unauthorized access. There are many approaches for securing the single-key-sharing transmission, but each has its own drawbacks. In this paper, we propose and present a novel approach to secure key-sharing over a communication network in symmetric key cryptography system which makes the single-key immune to unauthorized access. A total of four messages are exchanged between two devices for our secure key-sharing method. To implement the key sharing process, our method employs a few techniques, including asymmetric key cryptography, hash functions, machine learning-based pseudo random number generators, and timers. Our analysis shows that, besides providing similar level of confidentiality as the existing approaches, it also provides other significant improvements over the current ones, such as enhanced integrity maintenance, and authenticity verification of the two devices involved in the process. The short latency of modern 5G networks helps to balance the increased network demand caused by sending four independent messages. To determine timer duration and key validity, we propose applying AI algorithms and extending the security of our method; nevertheless, these applications fall within the purview of our upcoming study
A Study on Wearable Tech Interfaces and Perception: Cognitive AI-Enabled Device
Get PDFIn our paper, we explore the possibility of utilizing state-of-the-art hardware architecture to develop an interactive Artificial Intelligence(AI) based virtual agent in an Internet of Things (IoT) enabled smart glass. In contrast to the traditional system that relies on Central Processing Unit(CPU) and Graphical Processing Unit(GPU), we examine the possibility of a proposed hardware system that employs neuromorphic computing for precise and energy-efficient processing in real-time, resolving the problems of latency and excessive power consumption. When combined with conventional Complementary Metal-Oxide-Semiconductor (CMOS) technology, Resistive Random Access Memory(ReRAM) offers non-volatile, fast memory that facilitates parallel computing and guarantees smooth data store and retrieval for wearables with limited resources. This study also emphasizes the application of intelligent virtual agents based on cognitive AI in wearable technology. In order to create an immersive human-computer interface, we attempted to create an intelligent interactive AI avatar with Cycle Generative Adversarial Network (CycleGAN) that mimics the user\u27s traits and further pushes it to the large Language Model(LLM) to generate motion sequences that perform additional tasks, transforming LLM prompt reactions into movements.
The Evolution of Software Engineering: From Prehistoric Beginnings to the Age of Artificial Intelligence
Get PDFThis article presents a thorough overview of the historical background of software engineering, tracing its roots from antiquity to the contemporary time dominated by artificial intelligence (AI). Beginning with the basic programming languages of the mid-1900s, the article stresses major milestones, including the birth of object-oriented programming, the expansion of software engineering practices, and the transformational impact of structured programming. Additionally, the paper examines the limitations, possible hazards, ethical considerations, and environmental consequences of AI-powered software engineering, including machine learning (ML), the Agile revolution and its revolutionary impact on software development practices. Through the analysis of this developmental trajectory, the essay offers vital insights into the progressive nature of software engineering and the pivotal role of artificial intelligence in shaping its future
A Helical Model of Color Harmony
No full textAmong the many approaches to the study of color harmony tried so far, a relatively recent method is to leverage the large number of human-created and ranked color palettes, such as those hosted at colourlovers.com. Analysis of these large datasets could provide insights into the nature of color harmony. In this study, a large number of palettes with five colors were observed in 3D in different color spaces. It was found that a significant number of such palettes fit a single pitch of a helix aligned along the lightness axis, but not centered at the origin of the a-b plane in CIELAB and Oklab spaces. Considering the presence of an accent color, more than 50% of the highly ranked palettes studied fit the helical model. The helical model was then used to create new color combinations. In a survey, respondents were asked to like or dislike the patterns colored with these color combinations. It was found that the new color combinations thus formed were almost as harmonious and pleasing as the originals