LC International Journal of STEM (ISSN: 2708-7123)
Not a member yet
120 research outputs found
Sort by
Indoor Smoking Detection Based on YOLO Framework with Infrared Image
Get PDFThis study recommends combining the efficacy of YOLO with the greater visibility provided by infrared imaging to create a better indoor smoking detection system. The YOLO system divides photos into a grid and anticipates bounding boxes and class probabilities at the same time, making it an obvious choice for its real-time item detection capabilities. The approach improves its robustness by identifying heat signals associated with smoking sessions and overcoming limitations posed by low-light or blocked circumstances. The addition of infrared images significantly improved the system's performance in low-light conditions. A dual spectrum thermal camera is used in the entire indoor smoking detection system to obtain a large collection of infrared images representing various interior locations with documented smoking episodes. During the training phase, data augmentation processes such as random rotations, flips, and brightness and contrast fluctuations were used to improve the system's performance. The CIoU loss function improved the system's localization accuracy significantly, reducing false positives and improving overall detection performance. The combination of YOLO and infrared photography, in conjunction with data augmentation and the CIoU loss function, not only improves indoor smoking detection but also demonstrates the benefits of merging several technologies in the development of more effective and adaptive systems
Research on Intelligent Control of a 10-Channel Microwave Input Heating Microwave Oven
Get PDFThe increasing demand for precise temperature control and specialized process control in industrial microwave ovens has led to the exploration of advanced control algorithms. To address these challenges, innovative neural network control algorithms have been introduced. This article delves into the heating mechanism of a 10-channel high-power industrial microwave oven and offers a mathematical explanation for the microwave heating process in the chamber. Through MATLAB simulations, the heating process and the RBF neural network adaptive control system were investigated, demonstrating promising performance. An intelligent control system was then designed, incorporating components such as a 10-channel magnetron, microwave cavity, temperature sensor, and STM-32 microcontroller. Utilizing an RBF neural network adaptive control algorithm, this system independently adjusts 10 microwave inputs to achieve heating and maintain the desired temperature. Subsequently, a 10kW 10-channel high-power industrial microwave oven RBF neural network adaptive control system was implemented and experimentally validated for its effectiveness. This innovative approach offers adaptive intelligent control, enhancing performance across diverse operating conditions
Indoor Smoking Detection Method based on Dual Spectral Fusion Image and YOLO Framework
Get PDFIndoor fires are a major problem for public safety, with smoking being the most hidden threat. Traditional fire detection systems, such as smoke detectors, are only useful in the early stages and face challenges due to low light and limited visibility. This article describes an indoor smoking detection system that combines visible and infrared image fusion with the YOLO (You Only Look Once) detection framework. This technique improves indoor smoking detection performance by combining infrared thermal data with deep learning concepts. The YOLOv9 system detects indoor smoking behavior using a deep neural network for feature presentation and inference. The approach is optimized at the data, feature extraction, and model training levels to improve scene adaptability. The experimental results showed that on the custom indoor smoking dual spectral fusion image dataset, the average accuracy mAP (@ 0.5) of the Modified YOLOv9c detection model reached 95.8%, which was much better than the baseline models YOLOv5s (81.4%), YOLOv7 (89.7%), YOLOv8 (90.8%), and YOLOv9c (89.9%) mAP, respectively with significant performance improvements. Strategies like dual spectrum fusion, data augmentation, attention mechanism, and loss function were implemented to improve model detection performance. This paper presents a practical solution for indoor smoking detection tasks, demonstrating the approach's superiority in detection performance and providing a viable toolset for public safety against indoor fire hazards
The Impact of Automation and Artificial Intelligence on Employment Dynamics in Pakistan's Manufacturing Sector
Get PDFThe study titled The Impact of Automation and Artificial Intelligence on Employment Dynamics in Pakistan's Manufacturing Sector" was planned to understand the effects of automation and AI on employment in Pakistan's manufacturing sector. The investigation was concerned with the influence of a rising uptake of these technologies in job creation, skill demands, and labor displacement. It also evaluated various industries' technological preparedness levels for AI-led revolution and organizations' and policymakers' competitive and sectoral response. Through the statistical data on employment, the research determined the areas exposed to automation risks – however, the study only provided a math estimate of the number of jobs that could be potentially threatened by automation. The research also raised awareness of re-skilling and the creation of new competencies, especially in emerging technologies such as artificial intelligence, data science, robotics and machine learning. The research highlighted the need for policymakers to intervene, develop viable trade skills and talents for the workforce, and for businesses to get together to reduce job losses and benefit from the opportunities created by automation. The control group study also affirmed, thus, the need for a proactive approach to policy measures designed not only to address the social implications of automation or the changes in the occupational structure towards AI-intensive work but also to support the required strengthening of the links between skills development and the demands of modern manufacturing in Pakistan
Exploring Data-Driven Approach for Financial Fraud Detection: A Comprehensive Literature Review
Get PDFFinancial fraud detection has emerged as a critical area of research with the growing complexity and scale of fraudulent activities in the financial sector. Traditional methods of fraud detection, which are based on rule-based systems and manual oversight, fail to capture the dynamic and sophisticated nature of modern fraud schemes. This comprehensive literature review examines data-driven approaches that take into account the advancement of machine learning, artificial intelligence, and big data analytics to improve fraud detection. Some of the key methodologies covered are supervised, unsupervised, and hybrid models. The survey reflects growing usage in neural networks, ensemble methods, and anomaly detection techniques, emphasizing their performance in identifying complex fraud patterns in different financial datasets. Discussions include the difficulties with unbalanced datasets, evolving tactics for frauds, and requirements for explainability that remain future areas of interest. Drawing upon recent relevant research work, this review synthesis aims at informing readers concerning the landscape evolution in fraud detection against finances and presenting possible innovations in order for these to remain robust yet adaptive, clear, and transparent in nature
Programmer Assisted Tool Impact on Static Error Handling Capability of Novices in Imperative First Programming Languages
Get PDFLearning and understanding the syntax of a programming language is an extremely ordeal for novice programmers majoring in computer science. Introduction to programming is offered as a core subject. Novices have to use IDEs to write their programs. These IDEs has a valuable impression of novice error handling skills as static error messages are represented as intricate compiler waffles, terms and puzzling sentences. Docile, easy to understand, simple error messages are of prissy importance to evaluate novice programming aptitude. This research represents the outcomes of programmer assisted tool in natural language for explication of static errors in an imperative first programming language like C. Programmer Assistant tool (PAT) represents natural language description/solution of static errors underpinning Human Computer Interaction (HCI) approach, in the IDE and work as an offline static code analyzer. To assess effectiveness of this PAT novices was directed to write programs in different IDEs first and later using this PAT. Frequency of static errors, error, problem-solving time was analyzed and compared. The result of this study depicts that use of the programmer assisted tool has deep impact on novice performance, motivation and learning outcome. The quantitative mathematical analysis of our study revealed programming assistant tools has significantly influenced programming and static error handling skills of the students majoring CS
Distinguishing Between Drones and Birds Using CNNs Algorithm
Get PDFRecognizing drones or unmanned aerial vehicles (UAVs) from birds is a crucial capability for numerous applications. We create a convolutional neural network (CNN) drone identification system that can distinguish between images of drones and birds. A dataset of photos taken of birds and drones in various settings is used to train the CNN model. Our model distinguishes between drones and birds with 93% accuracy. The excellent results show that CNNs are capable of accurately differentiating between drones and birds under practical circumstances. Overall, this work demonstrates that deep learning may be used to achieve accurate drone recognition when similar avian items are present
Quality Status of Existing Services of Higher Education: A Study on Selected Govt Colleges of Northern Bangladesh
Get PDFThe right to education is the right not only to access education but also to receive quality education. Quality Services need to be assured in every level of educational institutions. It is very much needed for higher education as it supplies bottom to top most levels human capital to labor markets. Govt. colleges of Bangladesh affiliated to National University play a gearing role to increase the rate of higher education. Bangladesh education sector is realizing the extensive impact trying to ensure customer services with its tangible and intangible dimensions to increase the quality of education. Especially quality education of Bangladesh, is facing enormous challenges regarding quality of services provided by the institutions. Stakeholders (students and related others) do not get due quality services from the educational institutions. It is truer for higher education institutions like the government colleges. The study is trying to explore the present scenarios of services and their quality status by interviewing 80 (20 from each four selected government colleges of northern Bangladesh) teachers, staffs and service personnel through descriptive statistics by a structured questionnaire and a checklist. The study concludes to provide real scenarios about present status of services in government colleges. The study finds that among fourteen service dimensions most are average and below average and few service dimensions are in good status
Maintenance and other Operating Expenses (MOOE) and School Based Management (SBM) Performance of Secondary Schools in Samar Island
Get PDFThe study investigated the extent of Maintenance and Other Operating Expenses (MOOE) utilization relating to School Based Management (SBM) performance of public secondary schools in Samar Island for school years 2016-2019. A mixed method, explanatory research design was employed utilizing 159 school heads. Focus Group Discussion (FGD) was conducted for selected School heads, disbursing officers or bookkeepers and teacher representatives from six divisions in Samar Island. Quantitatively, this study made use of percentages, weighted mean, and Pearson r for the correlative data; while thematic analysis enriched the qualitative investigation. Findings of this study highlights a high extent of MOOE utilization to ensure students’ access to complete basic education; priority for school ceremonies like the Moving Up of the Junior High School, Closing Ceremonies, and Recognition Programs; to support the needs of the learners for their learning activities; to fund the rentals and minor repairs of tools and equipment that are important to classroom activities; and to fund the expenses related to graduation rites. The results showed that the least priority in the expenses were for the Internet connection of the school, telephone bills, salaries for the janitors, and the security guards. In addition, it was also found out that the SBM performance of the schools have improved in the past three years, from Good to Better. Finally, no substantial connection prevailed between the extent of utilization of MOOE and SBM performance of secondary schools in Samar Island
A New Ranking Technique to Enhance the Infection Size in Complex Networks
Get PDFDetecting the spreaders/sources in complex networks is an essential manner to understand the dynamics of the information spreading process. Consider the k-Shell centrality metric, which is taken into account the structural position of a node within the network, a more effective metric in picking the node which has more ability on spreading the infection compared to other centrality metrics such the degree, between and closeness. However, the K-Shell method suffers from some boundaries, it gives the same K-Shell index to a lot of the nodes, and it uses only one indicator to rank the nodes. A new technique is proposed in this research to develop the K-Shell metric by using the degree of the node, and a coreness of its rounding friends to estimate the ability of the node in spreading the infection within the network. The experimental results, which were done on four types of real and synthetic networks, and using an epidemic propagation model SIR, demonstrate that the suggested technique can measure the node effect more precisely and offer a unique ordering group than other centrality measures