Spectrum of Engineering Sciences

BLOCK CHAIN DRIVEN SUPPLY CHAIN SECURITY: INTEGRATING POST QUANTUM CRYPTOGRAPHY WITH AES

2025-06-02T11:25:48+03:00

Conventional cryptographic methods are gravely threatened by the development of quantum computing consequently, post quantum cryptographic (PQC) methods must be applied to secure data flow. In this work, we propose a hybrid encryption strategy based on Advanced Encryption Standard (AES) and Kyber, a lattice based PQC method, to raise the safety and processing efficiency of block chain systems, including multi stakeholders. Kyber codes keys Advanced Encryption Standard (AES) encrypts data. This approach lowers processing delay and is post quantum safe. Based on a performance study, the RSA encryption procedure takes 0.00049 seconds, and the decryption process takes 0.003051 seconds, therefore producing a total delay of 0.003524 seconds. Advanced Encryption Standard (AES) is quite efficient in generating a total delay of 0.000022 seconds, given a decryption time of 0.00001 seconds and an encryption time of 0.000012 seconds. The PQC hybrid (Kyber + AES) model helps one to reach these numbers with an encryption time of 0.0012 seconds, a decryption time of 0.0015 seconds, and a total processing delay of 0.002 seconds. This approach allows one to strike a compromise between real time communication effectiveness and safety. Research findings highlight the relevance of the concept in post quantum block chain systems, decentralized networks, and safe multiple party transactions, all settings where data integrity and secrecy are paramount. The paper presents a novelty in that it proved the existence of a hybrid encryption technique resistant to pragmatic and quantum computers.

CONNECTED REGIONS FORMATION FOR IMAGE CLASSIFICATION

2025-06-02T12:02:25+03:00

Many computer vision applications rely on matching key points between images. Over recent decades, advancements in key-point detection algorithms have significantly improved both robustness and speed. However, there is an ongoing need for more compact descriptors and faster methods with higher classification accuracy. This work addresses this need by introducing a novel algorithm that formulates both a key-point detector and descriptor based on prominent image features. The proposed detector focuses on identifying intensity-based corners and edges within grayscale images. This process involves detecting connected regions by analyzing pixel intensity ranges. Once bright and dark regions are identified, pixel intensities are sorted accordingly. Symmetric sampling is then applied after cascade matching, utilizing 128-bit descriptors. Isotropic and anisotropic filtering techniques are applied to the maximum filter response of the grayscale image. To normalize the descriptors, L2 normalization is performed on the RGB query image. The resulting feature vectors are spatially organized, and Principal Component Analysis (PCA) is applied to reduce their dimensionality. To improve search efficiency, indexing and searching are performed based on a visual words representation of the database of visual features. The proposed method was evaluated using two datasets, Caltech-256 and Corel-1000, and compared to the standard HOG detector and descriptor. Experimental results show significant improvements in both average precision and average recall for the proposed method.

AI-POWERED TOOLS FOR FASTER AND BETTER SOFTWARE DEVELOPMENT

2025-06-02T12:26:03+03:00

Artificial Intelligence (AI) has changed many fields, and making software is one of them. With more need for fast, good, and nice code, AI tools are being used a lot to help in different parts of building software, from making code and fixing it to finding mistakes and checking software. This paper looks at how AI tools affect the speed and quality of making software by conducting a test. We put AI tools into four main groups: code finishing, auto-checking, mistake finding, and project helping. A test was done with two groups of coders — one using old ways of working and the other using picked AI tools. Important measures, like code quality, time to make it, and mistake counts, were checked and compared. The results show that AI tools boost work speed and cut down on human mistakes, but problems like tool correctness and learning time are still there. These findings show AI can be a big deal in modern software building, giving useful ideas for coders and researchers.

DESIGN AND IMPLEMENTATION OF A HEAD-MOTION CONTROLLED ELECTRIC WHEELCHAIR FOR ASSISTIVE MOBILITY IN QUADRIPLEGIC PATIENTS

2025-06-03T11:10:59+03:00

Quadriplegic individuals, who have lost motor control in all four limbs, require specialized assistive technologies to achieve mobility and independence. This paper presents the design and implementation of a head-motion-controlled electric wheelchair, developed to enhance the autonomy of individuals with severe physical disabilities. The system supports two operational modes: manual and automatic, with the latter utilizing head tilt gestures to control movement. A head-mounted inertial measurement unit (IMU) detects head orientation (forward, backward, left, right), and a microcontroller translates these inputs into movement commands. The mechanical structure comprises a standard wheelchair retrofitted with DC motors and a gear reduction system, while motion control is achieved using a custom-built, relay-based H-bridge motor driver. To enhance motion detection accuracy, sensor fusion is performed using a Kalman filter. Experimental evaluation shows that the wheelchair maintains a stable forward velocity of approximately 0.5 m/s, with smooth bidirectional turning and high command recognition reliability. The paper details the system architecture, hardware and software integration, control strategy, and performance assessment. Potential future improvements include emergency stop features via GSM, variable speed control, health monitoring, alternate input methods, and obstacle detection. The results affirm that the proposed system is a viable, cost-effective assistive solution, significantly improving mobility and quality of life for quadriplegic users.

LINEAR REGRESSION MODEL IN CONTEXT OF MOBILE APPLICATIONS USAGE

2025-06-04T09:41:32+03:00

The popularity of mobile applications has resulted in an ever-increasing number of programmes being installed on smartphones. Whether or whether it is possible to predict which app a user will open is the subject of this study. The ability to forecast what apps will be needed in the future can aid in pre-loading the required apps into memory or in floating the relevant apps to the home screen to speed up launch times. We analysed a wide range of contextual information from the MDC dataset, including the user's profile, time and location, and the most recent App they utilised. The findings of our investigation can be divided into three categories. First and foremost, contextual information may be utilised to better understand how a user interacts with an app and to make more accurate predictions about how that app will be used in the future. A large part of the forecasting accuracy for the MDC dataset comes from the correlation between the sequentially used applications. The linear model is better than the Bayesian model because it can take into account all of the relevant information and provide more precise predictions than the latter. Predictions about app usage based on contextual information such as time, location and user profile and the most recently used app have been offered as a consequence of our research. For app usage prediction, we studied the topic of context awareness, and one of the things we observed was that context can significantly affect a user's app usage behaviour. We can deduce some patterns about how mobile app users interact with the software by examining contextual data. Personal mobile systems that employ contextual information to dynamically offer information, such as Apps to be used, to the user and improve the user-mobile phone interaction experience are suggested by this work. A personal mobile device is an example of this type of system.

REAL-TIME CRACK DETECTION IN MATERIALS USING A NOVEL CRACK-AWARE CNN-VIT HYBRID MODEL

2025-06-04T09:53:52+03:00

Undetected cracks in materials like concrete, asphalt, metals, and composites jeopardize structural integrity, posing safety and economic risks across infrastructure, aerospace, and automotive sectors. This study proposes a Crack-Aware CNN-ViT Hybrid model for real-time crack detection, integrating a Crack-Aware Attention Module (CAM) to emphasize crack geometry and a Crack Severity Annotation Framework to classify cracks by width, depth, and impact. Trained on a 60,000-image RGB dataset, augmented with conditional Generative Adversarial Networks for diverse materials and conditions, the model achieves 95.3% ± 0.2% accuracy, 94.2% ± 0.3% precision, 96.0% ± 0.2% recall, 95.1% ± 0.2% F1 score, and 90.5% ± 0.4% IoU at 32 fps, processing webcam feeds on an NVIDIA Jetson Orin Nano. Ablation studies, cross-dataset validation on SDNET2018 and CrackTree260, and a real-world bridge inspection demonstrate statistically significant improvements over YOLOv8 (by 5.1% accuracy) and Vision Transformers. Enabling automated, edge-based monitoring with timestamped crack storage, this scalable solution advances structural health monitoring, ensuring predictive maintenance and safety.

ARTIFICIAL INTELLIGENCE (AI)-POWERED PREDICTIVE MODELING FOR PATIENT READMISSION AND TREATMENT RESPONSE USING ELECTRONIC HEALTH RECORDS AND MACHINE LEARNING

2025-06-04T11:30:42+03:00

Predictive modeling using electronic health records (EHRs) and machine learning can revolutionize the medical field by exposing high-risk patients and refining treatment strategies. Objective: Using EHRs and machine learning techniques, this research seeks to develop and evaluate AI-powered predictive models for patient readmission and treatment response. Methods: Using a quasi-experimental study design, the effectiveness of AI-powered predictive models in projecting patient readmission and treatment response was assessed. There were 200 adults, age (≥18 years), in the sample. Carried out in a hospital setting, the study uses electronic health records (EHRs) and allows evaluation of AI-driven predictive algorithms in a real clinical environment. Electronic health records (EHRs) are the primary data source for the study. EHRs give extensive data on patient demographics, treatment outcomes, and medical history. Descriptive statistics; logistic regression; machine learning algorithms (random forest, support vector machine); model performance evaluation using metrics such accuracy, precision, recall, and area under the receiver operating characteristic curve AUC-ROC. Results: The model identified significant readmission risk factors with an 85% accuracy rate. Conclusions: By identifying high-risk individuals and fine-tuning treatment protocols, AI-powered predictive modeling has demonstrated its ability to improve patient outcomes.The findings suggest that clinical decision support systems providing personalized recommendations for patient care could be developed using artificial intelligence.

THE DENSITY IMPACT ON THE MATHEMATICAL MODEL OF THE SELF-PROPELLED PARTICLES IN THE HETEROGENEOUS MEDIA

2025-06-04T11:40:37+03:00

The collective dynamics of the self-driven particles was investigated mathematically and computationally. A two-dimension square shaped heterogeneous medium with boundaries was developed to study the behavior of the self-propelled particles in the existence of the static obstacles. It was examined that how changes in density impact on the collective motion of the SPP’s within the Heterogeneous media. The maximum cohesive motion and ferromagnetic alignment of the particles was determined in the presence of optimal noise . It was observed that as the number of the particles was increased, exceptional collective alignments among the particles were formed. At and an uninterrupted continuous escalation in motion was identified and the value of collective motion of self-propelled particles was reached at out of at time steps. The inherited order parameter among the particles in this model is an important critical factor.

UTILIZING ARTIFICIAL INTELLIGENCE AND DEEP LEARNING MACHINE-LEARNING APPROACH FOR OPTIMIZING DRUG DELIVERY SYSTEMS

2025-06-06T07:53:38+03:00

The thesis structure covers in short introducers the role of machine learning in explaining future behavior, and in depth it investigates the concepts of supervised, un-supervised, semi-supervised, reinforcement learning, highlighting deep learning. Hypertension is one of the most significant public health issues globally, with millions of individuals infected. Therefore, accurately predicting treatment groups for patients with hypertension will assist healthcare providers in making informed decisions that will enhance the outcome. However this study aims to create machine learning model capable of predicting the best treatment group for patients with hypertension based on demographic and clinical traits. A patient dataset was used comprising individuals diagnosed with hypertension and different machine learning models were evaluated. Findings from this study imply that machine learning models can be applied in predicting the ideal treatment group for hypertensive patients mandatorily. This research used a dataset available on Kaggle named “Hypertension Treatment Clinical Trial Dataset.

UTILISING A SOLAR ENERGY CALCULATOR FOR EFFECTIVE SOLAR ENERGY PLANNING AND EQUIPMENT SIZING

2025-06-06T08:13:18+03:00

The Solar Energy Calculator project is a groundbreaking initiative that makes use of technology to encourage the use of sustainable energy sources. By harnessing a suite of web development technologies, including HTML, CSS, JavaScript, PHP, and MySQL, this initiative has given rise to a cutting-edge tool for estimating solar energy production. The user interface design prioritizes accessibility and user-friendliness, incorporating features such as visual cues and intuitive input forms. The seamless integration of geographic and environmental data into a robust MySQL database ensures the precision of solar energy predictions. The thorough system testing process, which includes a variety of testing forms and constant monitoring of important indicators, has acted as a measure for the system's dependability and quality. This initiative is a demonstration of technology’s transformational potential in tackling urgent environmental issues and fostering a more sustainable future. With global energy demands on the rise and a growing imperative to transition towards renewable energy sources, the Solar Energy Calculator emerges as a timely and essential tool. It empowers users to make informed decisions regarding solar energy adoption, fostering a cleaner and more sustainable energy landscape. The Solar Energy Calculator reflects a dedication to environmental stewardship and a brighter future in addition to being a cutting-edge technology advancement.

TRANSITIONING PESHAWAR BRT FROM DIESEL-HYBRID ENGINES TO HYDROGEN FUEL CELLS: SOCIO-ECONOMIC IMPACT & FEASIBILITY

2025-06-06T08:22:21+03:00

Peshawar's Bus Rapid Transit (BRT) system holds great relevance in driving Peshawar transport, such a revolutionary system which has created thousands of jobs, and created more business opportunities than ever. But it needs to be improved with the hydrogen fuel cell technology (HFC), transition of Peshawar's Bus Rapid Transit (BRT) system from a conventional fuel transport to a futuristic one like that of Hydrogen fuel cell (HFC) will increase chance to adapt to a sustainable urban transport, reduce greenhouse gas emissions, and improve socio-economic elements in the region. As per the census carried out the Hydrogen Fuel Cell Electric Buses (HFCEBs) market value is estimated to be around $8.31 Billion, with an annual growth of 19.78% indicating its total market value to $20.49 Dollar by 2030. Transition of the Peshawar BRT systems to all Hydrogen Fuel Cell Electric Buses (FCEBs) will offer operational advantages to the transport infrastructure of the city, increasing the buses extended ranges to far side of the city, enhancing rapid refueling systems, all of which aligns with the demands for a high-frequency and swift BRT system. Implementation of Hydrogen Fuel Cell technology will enable BRT Systems to produce zero carbon emissions, help in decreasing air and water pollution, along with lowering the cost of operation for driving the BRT System.

A DETAILED ANALYSIS OF EMOTION RECOGNITION USING HUMAN FACIAL FEATURES IN INTELLIGENT COMPUTING SYSTEMS

2025-06-06T13:16:05+03:00

Emotion recognition through human facial features has emerged as a vital area of research in the field of intelligent computing systems, with broad applications in human-computer interaction, surveillance, healthcare, and user experience enhancement. This paper presents a comprehensive analysis of facial expression-based emotion recognition, focusing on its theoretical foundations, practical implementation, and integration into intelligent systems. The study explores the psychological models of emotions, particularly Ekman’s six basic emotions, and their physiological manifestations on the human face. It further investigates various computational techniques used to detect and classify emotions, including traditional machine learning algorithms such as Support Vector Machines (SVM), as well as advanced deep learning models like Convolutional Neural Networks (CNNs). Multiple publicly available datasets, such as FER-2013 and CK+, are examined to evaluate system performance and accuracy. The paper outlines a step-by-step pipeline for emotion recognition, encompassing face detection, feature extraction, classification, and post-processing. Special emphasis is placed on the role of data preprocessing, real-time performance, and generalization across diverse populations. Experimental results highlight the effectiveness and limitations of current techniques, with quantitative metrics provided to support the analysis. The study also discusses challenges such as variability in lighting, occlusions, subjectivity of emotional expression, and cultural differences. Finally, it outlines future directions, including the integration of multimodal data (e.g., voice, gestures), ethical concerns, and the potential for real-time deployment in adaptive intelligent systems. This detailed investigation contributes to a deeper understanding of how emotion recognition can be effectively modeled and utilized within the framework of intelligent computing.

A COMPREHENSIVE STUDY ON CYBER SECURITY THREATS AND PREVENTION MECHANISMS

2025-06-06T13:24:24+03:00

This article expresses emerging security threats and how we use different techniques to prevent them. As matter-of-fact previous research study shows just how to apply these security methods with different solutions but not provide future solutions as much as we need it in current time situation some of them explain universal security standards. This article contains future directions with emerging universal standards for cloud environment. The aim of writing this paper is to clarify the standards which help to protect from different threats like system data breaches, unauthorized access etc. cyber security is one of the leading process which help provide data and information protection and make sure to refrain from internet threats cyber bullying and online harassments. To control any type of internet security threats and vulnerabilities cyber security widely uses protraction measurements of every individual. Cyber securities reduce risk level based on different type’s situations like IOT weaknesses, IAM (Identify and access Management) and lie BYOD (Bring your own Device). In this study we purposing prevention techniques and also giving solution for prevention of common security threats.

QUANTUM-SAFE ENCRYPTION FOR CLOUD SERVICES: A NEW ERA OF DATA PRIVACY

2025-06-06T18:37:39+03:00

As significant advances are being made in quantum computing, existing cryptographic protocols for cloud security such as RSA ECC and AES are under thread from powerful quantum computing capabilities like such as Shor’s and Grover’s algorithm. The study, "Quantum-Safe Encryption for Cloud Services: A New Era of Data Privacy" examines the threats associated with existing limitations of encryption protocols at the advent of quantum years. This study analyzes the various methods of Post-Quantum Cryptography (PQC) that are available to protect against attacks from quantum threats including lattice-based, code-based, hash-based, multivariate polynomial, and isogeny-based methods. This paper also examines the framework for Quantum Key Distribution (QKD) protocols such as BB84 and E91 and the proposed umbrella methodology for cloud data protection by utilizing both PQC and traditional encryption. It provides examine a framework for the implementation of their methodology along with potential future directions including Blockchain-Based frameworks, integration of AI and standards-based protocols. This study emphasizes the urgent and immediate need for quantum-safe encryption, which is vital for the long-term confidentiality and privacy of cloud-based data services in future.

EVOLUTION OF INFORMATION SECURITY TOOLS TO PREVENT ID THEFT IN THE BANKING SECTOR OF PAKISTAN

2025-06-06T18:47:16+03:00

Purpose – Recently information breaches are increased in banking sector of Pakistan. The implementation old and non-effective Information Security Tools to Prevent Identity (ID) Theft pose significant risks to banking sector of Pakistan. The lack of appropriate evolution of these tools in banking sector causes of posing many vulnerabilities in information security infrastructure of and needs to be investigated. The purpose of this paper is to study and investigate the evolution of information security tools to prevent identity theft in banking sector of Pakistan. This paper also identifies the information security weakness in the existing banking infrastructure of Pakistan. Design/methodology/approach – A qualitative case study approach is used to conduct the research. Three case studies are conducted in different banks of Pakistan. The total number of one-to-one semi-structured interviews conducted was 31. A framework for Evolution of Information Security Tools to Prevent ID Theft in banking sector was proposed by extending the guiding framework for knowledge sharing processes for ID theft prevention within organizations proposed by Maitlo et al., (2019).

Findings – This research found that Banks need to upgrade their Information security infrastructure. As existing information security tools are not sufficient to prevent identity theft in banking sector of Pakistan. Therefore, it is required to volute the ID Theft Prevention, ID Theft Risk Assessment and ID Fraud Identification Tools. A Managerial Support for Information Security is also required. Practical implications – The research evaluates the information security tools for ID theft prevention in banking sector of Pakistan. It identifies the flaws in information security infrastructure in banking sector and provides the solutions to prevent ID theft. A framework for evolution of the information security tools used to prevent ID theft is developed to strengthen the information security infrastructure of banking sector. It guides the managers for effective use information security tools to empower the information security of banking sector. Originality/value – This study provides a new framework which was developed in the new context to evolute information security tools for ID theft prevention in the banking sector of Pakistan.

PERFORMANCE EVALUATION OF ELECTRICAL CHARACTERISTICS OF BIFACIAL HALF-CUT PERC MONOCRYSTALLINE PV MODULES UNDER OUTDOOR CONDITIONS IN NAWABSHAH CITY

2025-06-10T21:41:56+03:00

Electrical characteristics of both front and rearside of bifacial photovoltaic modules are normally examined at standard test conditions and rarely at outdoor environments. This work was carried out to inspect the electrical properties (front and rearside) of bifacial half-cut PERC mono-crystalline photovoltaic modules in Nawabshah city’s opened environment. This study measured the PV module's power output (P), open-circuit voltage (Voc), short-circuit current (Isc), maximum voltages (Vmax), maximum current (Imax). Weather parameters included global solar radiation (Ir), ambient temperature (Ta), wind speed (Ws), and relative humidity (Rh). A light meter (HD-2302) was used to measure the amount of global solar radiation that struck the site's horizontal surface, while PROVA AVM-05 was used to record the ambient temperature (Ta), wind speed (Ws), and relative humidity (Rh). Also via the PV analyzer PROVA-1101, the electrical properties of the PV modules both (front and rear) sides were recorded from 09:00 to 16:00 hours at an interval of 15-minutes. The average Ir, Ta, Ws and Rh were found 842.4W/m², 29.7°C, 3.0m/sec and 46.0%respectively during study period. The PV module's average Pmax, including both the front and back sides, was 334.6W, with the rearside alone producing 27.6W. It is established form the study that the front side produced around 91% of output power and the rest is contributed by its back-side during analysis period at the same operational and environmental conditions.

REDEFINING SMALL-SCALE ENERGY HARVESTING: PSO-ENHANCED MICRO-NOTCHED TURBINES AND RF RECTENNAS FOR AUTONOMOUS LOW-POWER DEVICES

2025-06-11T09:02:48+03:00

The significant strides made in science over the past century have led to the development of technical industries, primarily driven by advancements in miniature electronic circuits. The continuous decrease in power consumption is due to technology that enables electronic systems to be smaller. According to Intel, power management becomes more difficult as the size of electronic gadgets reduces. Several forms of energy are available that could supply power to low-energy applications. The creative notched turbines’ experimental setup presents ideal characteristics; hence, they are best for optimal power production. When the MiNT technology is introduced into a system to evaluate power delivery, as in this analysis, it becomes a unique system that can generate usable energy. The power supply tracking system of this developed energy-harvesting device relies on maximum power point tracking, a significant component of the implemented system. Environmental conditions affect the physical dimensions and performance of the productive power output of many energy-producing devices, imposing limitations that must be considered.

Investigators applied simulations and measurements to confirm the analysis process and results, which indicated a tug–nose micro–notch turbine system that enhanced the maximum power point tracking system (MPPT). An improved MPPT system utilizes PSO (particle swarm optimization) and the emulating resistor technique to reduce hardware size and increase power efficiency (PSO). This energy-collecting equipment features a micro-notch design, and a smaller version is equipped with an upgraded MPPT for all testing purposes. A PSO (particle swarm optimization) with a resistor emulation method enhances power efficiency and reduces the physical size of the maximum power point tracking system (MPPT). Research applications using the suggested control method can run successfully with as few energy sources as possible because this enhanced system and systems with low resistance load power it. The successful culmination of this research will enable many Energy and Power harvesting systems to improve their power management and control capabilities of their integrated circuits, thereby constraining end devices and application systems.

INNOVATIONS IN SEMICONDUCTOR FABRICATION FOR 3D INTEGRATED CIRCUITS: TOWARD COMPACT ARCHITECTURES AND HIGH-DENSITY INTERCONNECTS IN FUTURE ELECTRONIC SYSTEMS

2025-06-11T09:18:09+03:00

The increasing demand for higher performance, reduced power consumption, and compact form factors in modern electronic systems has catalyzed the development of three-dimensional integrated circuits (3D ICs) as a next-generation semiconductor solution. This paper explores recent innovations in semiconductor fabrication technologies that underpin the advancement of 3D ICs, emphasizing their role in achieving compact architectures and high-density interconnects. Fabrication methods such as through-silicon vias (TSVs), wafer-to-wafer and die-to-wafer bonding, hybrid bonding, and monolithic 3D integration are discussed in terms of their technical principles, advantages, limitations, and implementation challenges. Particular attention is given to critical issues such as thermal dissipation, inter-die alignment, yield improvement, interconnect density, and material compatibility factors that significantly impact the reliability and scalability of 3D ICs. The study also highlights the role of advanced materials, low-temperature processing, and heterogeneous integration techniques that allow for the vertical stacking of diverse components, including logic, memory, analog, and sensor layers within a single package. These developments are enabling more compact, energy-efficient, and functionally versatile electronic systems, with profound implications for applications in artificial intelligence, high-performance computing, data centers, mobile devices, and Internet of Things (IoT) ecosystems. Furthermore, the convergence of design automation tools, novel packaging strategies, and industry standards is accelerating the commercial viability of 3D ICs. The paper concludes by identifying emerging trends and future research directions that will shape the continued evolution of semiconductor fabrication, positioning 3D integration as a cornerstone of the next era in electronic system design.

AN EFFICIENT OFF-LINE HANDWRITTEN ENGLISH ALPHABET CHARACTER RECOGNITION BASED ON HIDDEN MARKOV MODEL AND DISCRETE WAVELET TRANSFORM

2025-06-11T09:32:15+03:00

Computational efficiency is a matter of great concern in state-of-the-art English alphabet character recognition systems. In this paper, nine state Hidden Markov Model (HMM) for character recognition has been presented. Alphabetical character images are being divided into nine blocks that corresponds to nine respective states of HMM. Corresponding local features of the character are being extracted by using geometric based feature extraction algorithm. Training of the HMM is done by means of the Baum-Welch algorithm. Computational cost of proposed model is minimized by employing Discrete Wavelet Transform (DWT) prior to other dimensionality reduction techniques. The recognition is performed using a Viterbi algorithm to perform best path search in combinations of various character models. Experimental results on handwritten English alphabet character databases demonstrate that recognition accuracy of proposed model is comparable to the existing techniques with reduced computational cost.

A PERFORMANCE ANALYSIS OF HYBRID SHE PWM VS. SHM PWM TECHNIQUES FOR HARMONICS CONTROL IN MULTILEVEL INVERTERS

2025-06-12T07:25:29+03:00

A distinctive Multilevel Inverter arrangement has been meticulously designed to reduce the number of power electronic components. This inverter system emphasizes effective harmonic control by utilizing a technique which is Hybrid Modulation that integrates the SHE-PWM for Elimination Pulse Width Modulation and the SHM-PWM for Mitigation Pulse Width Modulation. Utilizing Fourier series expansion, the paper formulates six nonlinear equations. The proposed topology generates a quarter-wave odd symmetry output waveform, eliminating even harmonics and exclusively producing odd harmonics. For half-wave, specific harmonics are targeted for elimination or mitigation while maintaining the fundamental harmonic at its maximum. To resolve the six nonlinear equations, an objective function is designed and subjected to constraints, utilizing a Genetic Algorithm to determine optimal switching angles. This investigation will use a Genetic Algorithm (GA) to identify the best switching angles for the suggested topology and modulation strategy. Extensive simulation analyses and comparing indicate that this method works better than conventional multi-level inverters at similar levels. Particularly, compared to previous multi-level inverters operating at the same level, the suggested design achieves much reduced Total Harmonic Interference.

RURAL HEALTHCARE MANAGEMENT FOR SINDH USING SMART TECHNOLOGIES

2025-06-13T07:26:38+03:00

A poor infrastructure combined with insufficient medical staff and minimal technological resources prevents rural areas in Sindh from obtaining quality healthcare services. A Rural Healthcare Management System (RHMS) uses mobile health units with telemedicine along with a digital data collection platform to respond to these rural healthcare issues. The system establishes its mission to deliver prompt quality healthcare services at reasonable prices to distant communities. RHMS implements community health workers together with smartphones and cloud-based data storage and centralized dashboards and monitoring capabilities for decision-making functions. This paper examines the systematic approach to design and implements the system alongside expected outcomes and challenges with future improvement goals.

REAL-TIME AGE AND GENDER ESTIMATION USING A FINE-TUNED DEEP LEARNING MODEL AND OPENCV

2025-06-13T07:36:58+03:00

In today’s digital era, automatic age and gender classification plays a vital role in various applications, particularly with the growing use of social media platforms. Despite recent advancements in facial recognition algorithms, analyzing real-world photographs continues to pose significant challenges. This study leverages Convolutional Neural Networks (CNNs) in conjunction with the Caffe deep learning framework and OpenCV to evaluate the accuracy of age and gender detection. By applying the Haar Cascade technique for initial face detection, the proposed model demonstrates improved performance in recognizing multiple faces within an image and accurately estimating their age and gender. The model was trained using both positive and negative facial image datasets, and its performance was thoroughly evaluated.

CFD ANALYSIS OF BIFURCATED ARTERY WITH VARIABLE PARAMETERS

2025-06-13T07:51:25+03:00

Cardiovascular disease (CVD) continues to be a top health concern worldwide, frequently developing when fat, cholesterol, calcium, fibrin and wastes collect on artery walls. If blockages remain untreated, they may result in heart attacks, strokes, hypertension and sudden cardiac death. The complex design of bifurcated arteries puts them at risk of both wrong flow and the formation of plaques. The goal of this study is to see how the angles at which arteries split influence blood movement within them using Computational Fluid Dynamics (CFD). It examines wall stress, velocity and pressure in the presence of both Newtonian and non-Newtonian pulsatile flow conditions. The bifurcated artery model was first developed with SolidWorks and then analyzed using Ansys Fluent to examine how the shape of the bifurcation (angles of 35°, 50°, 65° and 75°) influences wall shear stress, velocity and pressure. At the outset, simulations were carried out for steady-state situations, treating blood as Newtonian and applying the energy equation to keep body temperature at 37°C. Non-Newtonian behaviour was captured and blood’s ability to thin under shear stress was modelled, after which pulsatile flow was used to correctly represent the dynamic cycles of the heart. The results suggest that higher bifurcation angles are connected to lower wall shear stress, reduced blood flow, greater pressure and more flow disturbances. It becomes clear from pulsatile flow that the blood takes longer to accelerate and loses more energy as the angulation increases. Analysis using a mesh with sensitivity revealed that results are fully resolved at 0.1 mm or higher. The study points out the important effects of arterial walls and blood flow properties on blood circulation and gives useful information for improving cardiovascular tests, estimating disease development and planning better vascular treatment approaches.

INTEGRATING REMOTE SENSING AND DEEP LEARNING FOR AGRICULTURAL DROUGHT MONITORING

2025-06-14T11:51:09+03:00

An understanding of agricultural drought is critical in the management of agricultural lands especially in times of drought stress, which has adverse effects on agriculture. Management of drought, especially in its early stages, relies heavily on indices that are typically derived from onsite observations; thus coordination is often disadvantaged owing to the numerous factors that lead to drought. This study aims to improve the accuracy of monitoring drought conditions via deep learning techniques assisted by remote sensing data. A custom dataset of 544 JPG images (150x150 pixels) was compiled, consisting of 335 images representing drought conditions and 209 images without drought. Different deep learning architecture were executed employing varied layer and activation function configurations. It was observed that where models were developed employing multiple layers and using ReLU and Sigmoid activation functions, the accuracy obtained was as high as 97%. Emphasizing the gradual but progressive applicability of deep learning models for more efficient and forward-looking agricultural drought relying on satellite images. This mode of drought management is on the increase and enhances the overall effectiveness and viability of the agricultural drought monitoring systems.

OPTIMIZING CONGESTION CONTROL FOR QUALITY OF SERVICE (QOS) IN BANDWIDTH-CONSTRAINED WIRELESS NETWORKS

2025-06-14T12:27:14+03:00

Modern wireless networks typically operate on a best-effort service model, which, while able to support both real-time and non-real-time traffic, often falls short in ensuring the required Quality of Service (QoS) for real-time applications. Real-time applications, such as video streaming, voice over IP (VoIP), and online gaming, are highly sensitive to network conditions and require a predictable, low-latency environment to maintain performance. However, the best-effort model does not prioritize traffic effectively, leading to poor performance under high network load, with issues such as high jitter, excessive delay, and increased packet loss. QoS in wireless networks is traditionally assessed through performance metrics such as throughput, jitter, delay, and packet loss, all of which are crucial in determining the user experience in real-time applications. These metrics directly impact overall network efficiency and user satisfaction, with high delay or packet loss leading to degraded service quality, particularly for latency-sensitive applications. In this context, this study introduces a novel QoS framework tailored specifically for bandwidth-constrained networks, where managing limited resources is crucial. Instead of relying on the traditional approach of over-provisioning bandwidth, which can be inefficient and costly, the proposed model employs differentiated services combined with dynamic scheduling based on real-time measurements of incoming data rates and packet classification. By dynamically adapting the network's resource allocation to the changing traffic demands, the framework ensures that real-time applications receive the necessary priority, while non-real-time traffic is handled more flexibly. This results in a more efficient use of available resources, as bandwidth is allocated based on real-time traffic characteristics rather than fixed allocations. The framework incorporates an optimized queuing mechanism that prioritizes packets based on their type and current queue length, allowing for more accurate traffic management. This mechanism helps minimize delays for high-priority packets, such as those associated with real-time applications, while ensuring that lower-priority packets are processed appropriately without congesting the network. By reducing packet waiting times and minimizing the chances of packet loss, the approach significantly improves the QoS for real-time traffic, even in environments where bandwidth is limited. Furthermore, the model aims to minimize resource over-provisioning, which is a common issue in traditional network designs that often result in underutilized resources or excessive costs for provisioning higher bandwidth than necessary.

INTEGRATING ADVANCED DEEP LEARNING ALGORITHMS FOR CLIMATE SYSTEMS: ENHANCING WEATHER FORECAST ACCURACY, REAL-TIME CLIMATE MONITORING, AND LONG-TERM CLIMATE PREDICTIONS

2025-06-16T11:22:05+03:00

The integration of advanced deep learning algorithms into climate systems represents a transformative breakthrough in atmospheric science, significantly improving the accuracy and reliability of weather forecasting, real-time climate monitoring, and long-term predictive modeling. This study explores the deployment of state-of-the-art deep learning architectures including convolutional neural networks (CNNs), recurrent neural networks (RNNs), long short-term memory (LSTM) networks, and transformer-based models to process and interpret vast, heterogeneous datasets collected from satellites, sensor networks, and numerical climate simulations. By effectively capturing complex spatial-temporal patterns and nonlinear dynamics inherent in atmospheric and climatic processes, these algorithms address critical limitations of traditional physics-based models and enhance predictive capabilities across multiple timescales. The paper details the application of these deep learning methods in improving short- and medium-term weather forecasts, reducing prediction errors, and enabling dynamic adaptation to rapidly changing atmospheric conditions. It further highlights their role in real-time climate monitoring, facilitating early detection and classification of anomalies and extreme weather events with high spatial and temporal resolution. In addition, the research investigates the promising potential of deep learning to complement conventional climate models in long-term decadal climate predictions, addressing uncertainties and variability inherent in extended forecasts. Key challenges such as data quality, interpretability, computational resource demands, and integration with established meteorological and climate modeling frameworks are critically evaluated. This work emphasizes the necessity of interdisciplinary collaboration among climate scientists, AI researchers, and data specialists to develop transparent, reliable, and operational deep learning-enhanced climate systems. Ultimately, this comprehensive study demonstrates the profound impact of advanced deep learning algorithms in revolutionizing meteorological and climate sciences, enabling more precise, timely, and actionable insights essential for climate resilience, disaster preparedness, and sustainable environmental management in an era of accelerating global climate change.

CAREERLLAMA: AN AI-POWERED PERSONALIZED CAREER RECOMMENDATION SYSTEM WITH PSYCHOMETRIC AND SKILL GAP INTEGRATION

2025-06-16T13:30:21+03:00

Career decision-making has become an increasingly complex cognitive and informational challenge, driven by the rapid evolution of industry demands and job roles. In the absence of adaptive and personalized guidance systems, individuals are often left to make suboptimal career choices, leading to skill mismatches, occupational dissatisfaction, and underutilization of workforce potential. These inefficiencies not only hinder personal development but also negatively affect organizational productivity and economic resilience. This study presents an intelligent, context-sensitive career recommendation framework powered by the LLaMA large language model, designed to generate personal- ized career pathways. The system synthesizes multi-dimensional user inputs—such as educational background, acquired skills, cognitive preferences, and psychometric traits—to provide data- driven career recommendations. Furthermore, the framework identifies existing skill gaps and suggests targeted upskilling strategies. By incorporating machine learning and natural lan- guage understanding into career planning, the proposed model offers a scalable solution to the growing misalignment between workforce capabilities and evolving occupational demands.

IOT EFFECTIVENESS IN SUPPORTING ACADEMIC WORK AND ENHANCING LEARNING EXPERIENCES

2025-06-17T10:07:41+03:00

This paper aims to examine students' perceptions concerning the effectiveness of IoT in the form of academic work support from the use of IoT tools and the improvement of learning in the context of higher education. A cross-sectional research approach was employed and quantitative data were gathered using a structured questionnaire administered among undergraduate students at the University of Sindh, Jamshoro, Pakistan. The questionnaire measured dimensions such as characteristics of the task, characteristics of the technology, task-technology fit, and the effect of the technology on performance. Descriptive statistics were used to analyze data. The results showed students generally perceive that the use of IoT facilitates education tasks efficiently, such as real-time access to information, enhancing classroom interactions, and the ability to deliver personalized studies. Concern with data protection, technical issues, and unreliable access to IoT facilities across institutions was also noted. In connection with the findings, the study recommends strategic alignment of IoT with academic tasks, investment in digital infrastructure, and capacity building through awareness and training for effective and secure adoption in higher education.

INTERPRETABLE AI-BASED MATHEMATICAL MODELING FOR SOLAR POWER GENERATION AND LOAD FORECASTING IN DISTRIBUTED ENERGY SYSTEMS

2025-06-17T10:18:46+03:00

With the increasing demand for decentralized and sustainable energy solutions, the requirements of precise and explainable prediction techniques in distributed energy systems are becoming ever more important. This work introduces an interpretable AI-powered framework for forecasting solar power generation and short-term load forecasting based on mathematically justified models. With the use of solar panels having capacities of 3 kWh to 6 kWh, the research utilizes mean power outputs as inputs to a series of AI models with explicit equations, such as Linear Regression, Polynomial Regression, Decision Tree Regression, and Support Vector Regression. Each model is developed using transparent mathematical expressions for easy analysis and implementation. In addition, the framework incorporates real load profiles derived from the literature to test the models' ability in demand forecasting. Performance measures like Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and Coefficient of Determination (R²) are utilized to benchmark predictive performance. The findings validate that interpretable AI models not only make accurate predictions but also maximize model interpretability making them very well-posed to be used for real-time energy management in distributed power systems.

SOLAR FORECASTING FOR CROP CULTIVATION IN PESHAWAR AND MARDAN REGIONS: A COMPARATIVE ARTIFICIAL INTELLIGENCE ANALYSIS

2025-06-17T10:35:48+03:00

In the regions of Peshawar and Mardan, where sunlight has a major effect on crop production, precise solar forecasting facilitates optimization in agricultural activities. Predictive analysis was performed on a meteorological dataset employing Artificial Intelligence (AI) methods like Long Short-Term Memory (LSTM), Random Forest (RF), and Support Vector Regression (SVR). These methods are used to analyze the data and forecast solar irradiance. Performance of the models is evaluated based on measures like Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and R² values. Experimental results show that LSTM outperforms RF and SVR in terms of both precision and accuracy and achieves an RMSE of 12.45 W/m² for Peshawar and 14.32 W/m² for Mardan. This research aims to enhance solar energy prediction with AI to increase precision in agriculture for semi-arid areas.

EXPERIMENTAL INVESTIGATION OF THE EFFECTS OF PRESSURE ON THE MECHANICAL, THERMAL AND MICROSTRUCTURAL PROPERTIES OF CLAY-BASED GEOPOLYMERS

2025-06-17T10:53:23+03:00

This study reveals that compaction pressure has a strong and positive effect on the mechanical and microstructural properties of clay-based geopolymers. Using naturally available clays from Cherat and Gilgit-Baltistan, activated with a 10M sodium hydroxide solution, geopolymer samples were successfully developed and tested under controlled conditions. The compressive strength results showed a clear trend: as the applied pressure increased from 5000 to 24,000 pounds, the strength of the geopolymer also improved, reaching up to 48.10 N/mm². This strength is almost four times higher than that of traditional fired clay bricks commonly used in local construction. XRD analysis confirmed that pressure promoted the formation of new crystalline compounds such as natrosilite, sodium acetate hydrates, and sodium hydrogen silicate, which are associated with stronger geo-polymeric bonding. SEM images further supported these findings by showing a progressive reduction in porosity and an improvement in particle bonding and matrix formation as pressure increased. The overall performance of the samples prepared under higher pressure was clearly superior in terms of strength and structural integrity. These results show that applying compaction pressure is a practical and effective method to improve the quality of geopolymer materials without the need for high-temperature processing. In the future, research can be extended to study the long-term durability, shrinkage behavior, water resistance, and thermal insulation properties of these materials. Additionally, combining different clays, using alternative alkali activators, or applying pressure in combination with mild heat curing may further enhance their properties and promote their adoption in sustainable construction applications.

PERFORMANCE IMPROVEMENT OF SAC USING NEW EMD AND MD BASED 2D-OCDMA

2025-06-17T11:10:22+03:00

The adaptability of a passive optical network necessitates the employment of an appropriate multiple-access method that can offer the requisite transmission capacity in terms of data, reach, and number of users while being simple to construct and low in cost. Because of its asynchronous nature and simultaneous access to the channel for different users, the spectral amplitude coding- photonic code division multiple access system is expected to offer the needed capacity. However, in high cardinality systems, the 1D character of the spectral amplitude coding method restricts the reduction of multiple access interference and the related phase-generated intensity noise. In addition, dividing available spectral windows restricts support for large cardinality in zero or fixed-phase cross-correlation systems. As a result, for high transmission capacity and a large number of users across a long distance, a new dimension must be added to the existing 1D code. To include spatial encoding, currently employed 2D spectral amplitude coding-optical code division multiple access systems use spectral/spatial coding techniques that necessitate a significant number of optical fiber media between the transmission and reception modules. This severely affects the practicality of implementing 2D optical code division multiple access in a passive optical network. As a result, spectral/temporal coding has been optimized for low-cost passive optical networks. To support a high number of users while maintaining a low bit error rate, spectral/temporal coding is necessary. Based on current 1D multidiagonal and improved multidiagonal codes, this work proposes a unique 2-D spectral/temporal coding method. We do a thorough mathematical study using bit error rate, quality factor, and eye diagram as performance metrics. The system's performance shows that the code is efficient in terms of user count, multiple access interference, and encoder/decoder architecture.

OPTIMIZING DAYLIGHT USING PASSIVE STRATEGIES: LIGHT SHELVES AND SOLAR TUBES

2025-06-18T08:05:02+03:00

This research investigates how well- advanced daylighting techniques work in atrium buildings, specifically looking at how light shelves and solar tubes can be used together. The importance of daylighting in sustainable building design cannot be overstated, as it provides advantages like better visual comfort, increased occupant health, and lower energy usage. Atrium buildings, featuring spacious central areas, are perfect for integrating these tactics, offering better airflow, more sunlight, and improved spatial connections. The main goal of this study is to boost natural light, decrease dependence on artificial light, and enhance energy efficiency. The performance of these technologies was evaluated through detailed simulations and case studies using advanced daylighting software and real-world assessments. Light shelves raised levels of natural light by 15% to 17% in areas with windows, whereas solar tubes increased brightness from 100 Lux to 200 Lux in areas without windows. Out of the various dimensions examined, 24-inch solar tubes exhibited the greatest uniformity and brightness levels. The combination of light shelves and solar tubes showed noticeable enhancements in visual comfort and energy efficiency, indicating that integrating these daylighting solutions can greatly improve indoor lighting conditions and support sustainable building practices. The results highlight the significance of creative daylighting methods in contemporary architecture, stressing their involvement in establishing healthy, efficient, and productive spaces. This research addresses a significant gap in Pakistan's educational building design by offering simulation-based daylighting solutions customized for local climatic settings and construction techniques. This study offers important information for architects, engineers, and building designers who want to enhance natural lighting and sustainability in built environments.

A HYBRID LIGHTWEIGHT AND EXPLAINABLE FEDERATED LEARNING MODEL FOR REAL-TIME INTRUSION DETECTION IN RESOURCE-CONSTRAINED IOT ENVIRONMENTS

2025-06-18T08:31:09+03:00

In recent years, the increasing adoption of Industrial IoT and smart infrastructure has created new challenges in cybersecurity, particularly concerning data privacy and real-time threat detection. This study proposes a lightweight Federated Learning (FL)-based Intrusion Detection System (IDS) that collaboratively trains a neural network model across distributed clients without requiring centralized data collection. Using the UNSW-NB15 dataset—a comprehensive benchmark for modern network threats—we simulate federated training across multiple clients using a compact feedforward neural network. The model is trained locally on each client and updated globally using the Federated Averaging (FedAvg) algorithm. Our approach preserves data privacy while maintaining high detection accuracy.

A COMPARATIVE ASSESSMENT OF SUPPLY CHAIN EFFICIENCY: BLOCKCHAIN-BASED VS. TRADITIONAL MODELS IN AGRI-FOOD SECTOR

2025-06-15T20:37:04+03:00

The food value chain in agriculture has great significance with the supply of nutritious, accessible, healthy, and appropriate produce, feed, fiber, and fuel. The food supply chain system is a field of considerable significance as it provides customers with inexpensive, nutritious, and adequate bread and butter. The use of modern techniques are essential to ensure the smooth operations of these value chains. This study has presented comparative analysis between traditional and blockchain-based food supply chain system. Positivist philosophy has been used in this research because the sample population of research has been set the inclusion criteria as published findings obtained from SCI journals and books from credible research about blockchain-based food supply chain and other systems rather than blockchain. Data has been gathered by using the secondary data collection technique. The deductive nature of research has been used the grounded theory analysis to analyze the collected data in systematic methodology. And apply descriptive statistical and inferential statistical analysis (independent sample t-test) with the help of SPSS v 26. Due to the good performance and features of Traceability, credibility, integrity, and sustainability a blockchain-based food supply chain is most efficient than a traditional system. Results have monitored that a blockchain-based system has been used to fulfill the modern need of the supply chain in the future.

EXPERIMENTAL STUDY OF THE EFFECT OF WATER-REDUCING ADMIXTURE AND RECYCLED FINE AGGREGATE ON MECHANICAL PROPERTIES OF CONCRETE

2025-06-18T12:49:42+03:00

The utilization of recycled fine aggregate (RFA) in concrete has gained significant attention due to its potential to reduce environmental impact and conserve natural resources. However, the incorporation of RFA often leads to a reduction in the mechanical properties of concrete. This study investigates the combined effect of water-reducing admixture (WRA) and RFA on the compressive and tensile strength of concrete. The experimental program involved the preparation of concrete mixtures with varying proportions of RFA (0%, 50%, 75%, and 100%) and after trail for different dosages of WRA (2%, 3%, and 4% by weight of cement) and WRAs give good workability by use of 4% of WRAs. Compressive and tensile strengths were evaluated at 28 days. The results indicate that the addition of WRA significantly improves the workability and mechanical properties of RFA concrete. The optimal combination of 75% RFAs and 4% WRAs yielded the highest compressive and tensile strengths, demonstrating the potential for sustainable concrete production without compromising performance.

EXPERIMENTAL AND NUMERICAL STUDY ON THE RISE OF BUBBLES IN THE VERTICAL WATER CHANNELS

2025-06-18T13:02:54+03:00

This work investigates numerically and experimentally the behavior of air bubbles rise i.e. path and motion in vertical water channels using a submerged needle and the impact of water channel size on the behavior of rising air bubbles and contrasts this current threedimensional (3D) study with the earlier 2D studies. Experimental investigations were taken by installing two high-speed digital cameras (HSC’s). For smaller flow rates, the bubble's trajectories have been observed to spiral rather than rectilinear. The volume of fluid (VOF) approach is used for numerical analysis to investigate the trajectory of air bubbles in different water channels. The bubble trajectories of the front (x-y) plane and superimposed view of the (y-z) plane recorded during experimental investigations are drawn using MATLAB. Then graphical representations are drawn in MS EXCEL. The bubbles rise and follow a spiral path. Water channel size does not affect the behavior of rising air bubbles.

CARBON CAPTURE AND UTILIZATION: DEVELOPING TECHNOLOGIES TO CAPTURE AND UTILIZE CO₂, SUCH AS CARBON CAPTURE AND STORAGE (CCS) OR CARBON UTILIZATION: CASE STUDY OFNORWAY, CANADA AND USA

2025-06-18T13:14:41+03:00

Increasing levels of carbon dioxide (CO₂) in our atmosphere from what people do is the main factor leading to climate change. For this reason, greater consideration is being given to CCU technologies in the ongoing aim to cut global greenhouse gas emissions and move to a reliable low-carbon economy. CO₂ is first taken out of industrial and energy-related processes and is then converted into valuable goods or kept for a long time. Such strategies cover methods called CCS and several ways to use carbon which can each help lessen climate change and help industries become more sustainable. Many people separate carbon capture into pre-combustion, post-combustion and oxy-fuel combustion types. The main reason post-combustion capture is widely used in industries and power plants is that it fits well with current equipment. CO₂ that is captured can either be locked up in underground places such as depleted oil and gas wells and deep salty underground water (CCS) or be used to support oil and gas production (EOR) as well as the making of synthetic fuels, polymers, concrete and various other chemicals. Using carbon to produce chemicals and fuels has gained interest because it helps limit greenhouse gases and creates products that can be sold. The recent progress in chemical engineering, materials science and catalysis has helped make the capture and conversion of carbon dioxide both faster and more cost-efficient. Using advanced sorbents, membranes and solvents has led to better collection of CO₂ and promising new electrochemical and photochemical routes are coming up for converting it. Even so, the use of CCU is not without problems, since it requires a lot of money, energy, new infrastructure and the outcome of CO₂-derived products is not clear. Life cycle assessments (LCAs) are important so that businesses can confirm that using carbon is beneficial to the environment and does not cause environmental damage in other parts of the supply chain. For these barriers to be overcome, it is necessary to have policy help, team up with private companies and invest in specific research projects. Governments and such organizations are paying more attention to CCU as a way to achieve zero net emissions, as demonstrated by the U.S. Department of Energy’s Carbon Negative Shot and the Green Deal launched by the European Union. Additionally, uniting CCU technology with energy sources like solar and wind is likely to help industrial processes produce less carbon and contribute to a renewed carbon economy. All in all, using carbon capture and utilization technologies could solve climate change while ensuring industry is sustainable. More innovation, suitable policies and teamwork on a global scale help make CCU play its intended role in addressing climate change. Carbon capture and utilization (CCU) involves turning CO₂ into goods, resources or energy and is known as CCU.

THE ASSESSMENT OF THE IMPLEMENTATION OF BUILDING INFORMATION MODELLING (BIM) AND INTEGRATED PROJECT DELIVERY (IPD) IN THE CONSTRUCTION INDUSTRY: A CASE STUDY IN PAKISTAN

2025-06-19T08:25:33+03:00

This research explores the implementation and impact of Building Information Modelling (BIM) and Integrated Project Delivery (IPD) within Pakistan’s construction industry, examining how these methodologies could enhance project outcomes. Utilizing a quantitative survey, the study gathered data from industry professionals to evaluate the extent of BIM and IPD utilization, assess perceptions, and identify barriers to implementation. The results reveal that BIM adoption in Pakistan is notably low, with an estimated implementation rate of just 11% nationally, while findings from this study indicate an even lower rate among surveyed firms. Key challenges identified include a significant shortage of skilled professionals, insufficient training opportunities, and the high initial costs related with BIM technologies. Despite a growing awareness of the benefits of BIM and IPD, substantial gaps remain in their practical application, with many professionals still unfamiliar with or resistant to these modern methods. A comparative analysis with global adoption rates underscores a stark disparity between developed and developing nations. While developed countries have seen widespread integration of BIM and IPD, driven by government mandates and advanced infrastructure, developing countries like Pakistan face considerable hurdles. These include a lack of standardized practices, inadequate regulatory support, and limited resources for training and development. The study concludes that addressing these barriers is critical in maximizing the potential of BIM and IPD in Pakistan. Government support, through policy initiatives and incentives, combined with enhanced educational and training programs, is essential for overcoming the current obstacles. By focusing on these areas, Pakistan can fully leverage BIM and IPD to enhance efficiency and support the overall economic growth of the construction sector.

OPTIMIZING RESOURCE ALLOCATION AND SCHEDULING STRATEGIES IN SOFTWARE PROJECT MANAGEMENT: A SYSTEMATIC APPROACH

2025-06-19T13:23:25+03:00

Efficientresourceallocationandschedulingarepivotalinsoftwareprojectmanagementto optimize performance, reduce costs, and ensure timely delivery. This paper explores methodologiesandframeworksthatenabledynamicresourceallocationandadaptivescheduling. By integrating predictive analytics, workload distribution strategies, and cost-conscious resource provisioning, project managers can balance competing demands of scope, budget, and timeline. in today's fast-paced development environment, static resource allocation methods often fall short in addressing real-time challenges such as sudden shifts in project scope or resource availability. To bridge this gap, dynamic and predictive approaches utilize advanced tools like machine learning and real-time data integration. These methods empower managers to make informeddecisions, minimizebottlenecks, and alignprojectexecutionwithoverarchinggoals. theproposedframeworkemployshistoricaldataanalysisandmachinelearningmodelsto enhance decision-making in resource distribution and scheduling. Empirical validation demonstrates its ability to minimize delays and boost productivity when compared to traditional methods. Future directions include incorporating agile principles and real-time monitoring to further refine resource management practices, ensuring organizations remain competitive in an ever-evolving technological landscape. This paper explores methodologies and frameworks that enable dynamic resource allocation and adaptive scheduling. By integrating predictive analytics, workloaddistributionstrategies,and cost-consciousresourceprovisioning, projectmanagerscan balance competing demands of scope, budget, and timeline. The proposed framework employs historicaldataandmachinelearningmodelstoenhancedecision-makinginresourcedistribution and scheduling, ensuring alignment with project objectives. Empirical validation demonstrates that this approach minimizes delays and enhances productivity compared to traditional static methods. Future directions include incorporating real-time monitoring and agile frameworks to further streamline resource management processes.

ADVANCED CLAY-BASED GEOPOLYMER CEMENTS: STRUCTURAL BEHAVIOR, MATERIAL PROPERTIES, PARAMETRIC INFLUENCE, AND EMERGING APPLICATIONS

2025-06-19T13:35:29+03:00

The global construction industry is undergoing a transformative shift toward sustainable and low-carbon alternatives to traditional Portland cement. Among emerging materials, clay-based geopolymer cements have gained significant attention due to their eco-friendly synthesis, utilization of naturally abundant aluminosilicate clays, and exceptional mechanical and durability performance. This study provides a comprehensive investigation into the structural characteristics, physico-chemical properties, and influential synthesis parameters of clay-based geopolymer cements. Emphasis is placed on the reactivity of various natural clays such as kaolinite, montmorillonite, and halloysite under alkali activation, along with their phase transformations, microstructural evolution, and resulting geopolymer gel formation. Key parameters, including the Si/Al ratio, curing temperature, activator concentration, type of alkaline solution, and liquid-to-solid ratio, are critically analyzed for their impact on setting time, compressive strength, porosity, and thermal stability. The influence of calcination temperature and pre-treatment methods is also examined, particularly in enhancing the amorphous phase content and facilitating dissolution kinetics. Analytical techniques such as X-ray diffraction (XRD), scanning electron microscopy (SEM), Fourier-transform infrared spectroscopy (FTIR), and thermogravimetric analysis (TGA) are employed to characterize the mineralogical, morphological, and thermal behavior of geopolymer matrices. Furthermore, this paper evaluates the long-term performance of clay-based geopolymers in aggressive environments, including acidic and sulfate-rich conditions, highlighting their superior resistance compared to conventional cements. In addition to their structural utility, the multifunctionality of these materials is explored through applications in thermal insulation, fire resistance, carbon sequestration, and hazardous waste immobilization. By integrating original experimental findings with an extensive review of current literature, this study advances the understanding of clay-geopolymer chemistry and presents a pathway toward scalable, circular-economy-driven, and low-carbon construction materials for a sustainable future. The outcomes of this research are anticipated to significantly influence material selection in eco-construction, infrastructure resilience, and green architecture.

PREVALENCE OF DEPRESSION AND ITS ASSOCIATED FACTORS AMONG HIGH SCHOOL TEACHERS IN DIR(L) KPK, PAKISTAN

2025-06-20T10:54:48+03:00

In the present study, the prevalence of depression and its associated factors among high school teachers in district Lower Dir, KPK is investigated. Cluster random sampling is used to identify the representative sample from the population. A total of 180 teachers are selected from the study area. Their depression level is measured by scale (Centre for Epidemiological Studies Depression Scale Short Form, or CESD-10). A structured questionnaire is used for data collection. The data is presented by frequency distribution. The linear regression model is applied to assess the dependency of prevalence of depression on its risk factors. The model identifies only one variable out of ten (gender, age, residence, education level, income, marital status, chronic illness, workload, family support) variables, family support which is responsible for prevalence of depression. This indicates that in the study area, lack of family support is the major risk factor for incidence of depression.

PREDICTIVE REGRESSION MODEL FOR SUBSCRIPTION FORECASTING THROUGH TELEMARKETING

2025-06-19T20:32:37+03:00

n this study, a data mining technique was employed to predict the success of telemarketing calls aimed at promoting long-term bank deposits, a key challenge for financial institutions aiming to optimize their marketing strategies. The dataset used for this analysis was sourced from a Portuguese retail bank, containing 45,211 records and 17 different attributes, including demographic and behavioral information about the clients. The primary objective was to develop a predictive model that could accurately determine whether a client would subscribe to a term deposit based on the features available in the dataset. To build the predictive model, logistic regression was applied, a statistical method well-suited for binary classification tasks like this one. The model was trained to estimate the probability that a client would respond positively to the marketing campaign. A crucial step in improving model performance was the feature selection process, where 22 distinct sets of features were evaluated. This helped identify the most relevant attributes that contributed to the prediction, ensuring that the final model was both efficient and interpretable. The final model achieved a precision of 0.74 and a recall of 0.74, indicating that it performed well in both identifying positive responses (precision) and capturing as many positive responses as possible (recall). Specifically, the model made 11,294 correct predictions, including 6,124 true positives and 5,170 true negatives, demonstrating its ability to accurately classify both successful and unsuccessful subscription cases. However, it also made 4,047 incorrect predictions, which consisted of 2,505 false positives and 1,542 false negatives. These errors reflect the inherent challenges in predictive modeling, particularly in distinguishing between clients who are likely to subscribe and those who are not.

Keywords

(Logistic Regression, Predictor, Bayes, Telemarketing Success, Customer Data Analysis, Socio-economic Factors, Deep Learning).

MULTICLASS SKIN CANCER CLASSIFICATION USING RESNET AND DERMOSCOPIC IMAGING

2025-06-20T14:59:32+03:00

Skin cancer is among the most common and life-threatening illnesses globally, and early and precise diagnosis is necessary to enhance the outcome of treatment. Conventional diagnosis is based largely on skilled dermatologists, which is time-consuming and variable. Over the past few years, deep learning methods have proven to have vast potential in computerizing medical image analysis, both speeding up and improving the accuracy of diagnosis. This work aims to develop a reliable deep learning-based classification model using the ResNet architecture for identifying and categorizing various types of skin cancer. The research makes use of the HAM10000 dataset, an extensive collection of dermoscopic images representing various skin cancer types, to ensure representative and varied data for training and testing. The methodology suggested here encompasses several stages such as dataset preprocessing, feature extraction based on pre-trained deep learning models, and classification based on ResNet. Several data augmentation methods and preprocessing operations are utilized to improve model accuracy and handle imbalances in classes. Feature extraction takes advantage of ResNet's hierarchical feature representation, where complex patterns in dermoscopic images are captured to support precise classification. The model is optimized and fine-tuned with the best hyperparameters to achieve maximum classification accuracy with minimal computational complexity. A comparative analysis with traditional machine learning methods underscores the superiority of deep learning techniques in dealing with sophisticated visual data. Experimental findings show that the model proposed has excellent classification accuracy, performing better than traditional methods in precision, recall, F1-score, and overall classification accuracy. The model also separates various types of skin cancers correctly with minimal false positives and negatives. Performance monitoring methods like cross-validation and regularization are also embedded to prevent overfitting so that the model is generalizable to unseen samples. The computational cost of the method is examined to ascertain its viability for real-time clinical use, highlighting its prospect of being part of automated diagnosis systems.

Keywords

(Deep Learning, Skin Cancer, Skin Cancer Classification, Skin Images).

THE ROLE OF EXPLAINABLE AI IN MACHINE LEARNING MODEL INTERPRETABILITY

2025-06-20T15:30:49+03:00

This work investigates the contribution of Explainable AI to the interpretability of ML models by analyzing several methods of enabling model interpretability and how this benefit stakeholders through increased trust and usability. Explainable Artificial Intelligence has become an important part of the research area in machine learning to cope with the diamond's black box of verbose models. ML applications increasingly target sensitive sectors like healthcare, finance and law enforcement. It is making transparency and interpretability critical for building trust, enhancing decision-making and meeting regulatory requirements. This study is carried out using the mixed method, adopting the systematic literature review method alongside an empirical analysis of explainability techniques. A case study is performed in a real-world application, involving user perceptions and model performance trade-offs when using XAI methods. The discoveries clarify that while explainable artificial intelligence has procedures expanded the interpretability of a model. There was commonly a compromise between precision and explicability. This work highlights that the choice of explainable artificial Intelligence has method driven by the needs of the use case and goals of stakeholders. The task-specific efforts in developing scalable, such as, consistent, explanatory and real-time applicable. Explainable Artificial Intelligence has techniques are essential to promoting even wider integration of XAI methodology in ML-driven decision-making systems.

Keywords

Explainable AI , Machine Learning Interpretability, Model Transparency, Feature Importance, Trust in AI, Black-Box Models, Rule-Based Explanations, Decision-Making

OPTIMIZING DELIVERY ROUTES AND PACKAGE ALLOCATION FOR ENHANCED LOGISTICS EFFICIENCY

2025-06-17T12:30:50+03:00

With rapid advancements in e-commerce, increasing customer demands have put effective delivery management at the forefront of issues related to logistics and transportation companies. This study examines the Vehicle Routing Problem (VRP) and proposes a clever and metaheuristic way for effective delivery routing and distribution of items. Essentially, a means to improve scheduling, cost, resource productivity, and customer satisfaction. Research includes data on where items are delivered, where the depots are situated, what maximum loading each vehicle is capable of carrying, and obstacles to delivering services. Routing through clustering organizes the delivery points in respective modes and places where and how goods are to be delivered. Encoding uses ant colony optimization (ACO) to optimize routes through complicated problems such as vehicle constraints, time delivery, and dynamic destinations. Improvements in the algorithm performance shall also be attained by incorporating heuristics and evolution strategies. A web API was developed. During analysis, access to the most up-to-date routing information was ensured in real-time, with the application presenting the routes on interactive maps. Whereas static orders are currently managed, planned enhancements will allow for dynamic order handling, thereby enabling demand prediction via machine learning and real-time traffic information to enhance operational efficiency.

ARTIFICIAL INTELLIGENCE AND 6G INTEGRATION: TRANSFORMING THE DIGITAL TECHNOLOGY LANDSCAPE

2025-06-20T16:20:29+03:00

The combination of AI with 6G wireless networks is expected to transform the future digital era by facilitating novel advances in connectivity, automation, and intelligence. As research into 6G speeds up with the goal towards anticipated deployments around 2030, AI has become one of the cornerstones that move networks from static infrastructure into dynamic, self-optimizing ecosystems. This paper examines the synergistic opportunities between AI and 6G, including real-time data-processing, AI-native system architectures, and intelligent radio frequency beam forming, which could deliver terabits per seconds speeds, sub-millisecond latencies, and truly global coverage. Beyond such technical enhancements, the integration of AI and 6G will revolutionize industries such as healthcare, smart cities, autonomous transportation and immersive media by bridging the physical, digital and virtual worlds, for example through holographic communications, remote robotic surgery and the realization of real-time digital twins. But this evolution will not be without complications, from cybersecurity risks to expensive infrastructure and ethical quandaries around AI governance. Institution and media 5 pressures we believe that these may be addressed under a collaborative process that integrates efforts from academia, industry and policy makers, to ensure them and secure deployment. The merging of AI with the 6G would not just be an evolutionary development in wireless technology but also a paradigm shift to cognitive networks, which are able to make decisions by itself and learn from it. As global efforts struggle to define the contours of a 6G future, this paper calls for a balanced approach to innovation that prioritizes sustainability, inclusivity, and ethical frameworks to realize the untapped potential of intelligent connectivity. The age of AI-powered 6G does not just mean connecting the world more seamlessly, but redefining the very texture of digital interaction for decades to come.

Keywords

(6G networks, Artificial intelligence, Network optimization, Smart cities, Cybersecurity).

Next-Gen Customer Retention: A Stacked Ensemble Model for Churn Prediction

2025-06-20T18:16:56+03:00

Customer churn is one major problem in the telecom industry, requiring efficient and effective predictive models for proactive customer retention. Much work has already been achieved in this direction, but most studies so far have focused mainly on individual classifiers. While these models perform well in many areas, they have their own set of weaknesses. These include computational inefficacy, susceptibility to dataset imbalance, and inability to learn from subtle relations. Therefore, they tend to be insufficient in optimizing the trade-off between computational cost and accuracy. Existing methods tend to be insufficient with high-dimensional datasets, vulnerable to overfitting, or non-generalizable across telecom datasets. This work proposes an Ensemble Stacking model that is capable of overcoming these weaknesses. The proposed model consists of a set of base learners, Random Forest, Naïve Bayes, and K-Nearest Neighbors that are responsible for learning various patterns in the dataset. The base models make predictions, which in turn feed a single meta-model, Logistic Regression, that learns from their predictions to make the final prediction. The results reveal that the proposed model is capable of generating excellent accuracy with acceptable latency, outperforming all individual classifiers. Its superior latency-aware accuracy Index (LAAI) score also validates the fact that it is highly robust and adaptable, making it a very effective solution for real-world prediction problems.

Keywords

Customer churn prediction, Ensemble stacking model, Machine learning classifiers, Proactive customer retention, Latency Aware Accuracy Index (LAAI)

HAND GESTURE TO VOICE CONVERSION USING ARTIFICIAL INTELLIGENCE

2025-06-21T11:32:43+03:00

Communication forms one of the basics of interaction among people but becomes a significant problem for deaf and mute individuals. Conventional approaches, such as sign language and lip reading, tend to be restricted regarding availability and precision. The proposed project will close this communication gap with the help of an image-processing-based hand gesture recognition system. This system allows hand signals to be captured by a webcam, converted to text, and then to speech, providing a readily understandable input/output medium. This work captures the hand gestures, preprocesses the model to be invariant to lighting and background variations, and tests with measures such as IOU (Intersection Over Union), MAP (Mean Average Precision), MAE (Mean Absolute Error), and RMSE (Root Mean Square Error). Further, the proposed approach provides better results for hand gestures using image processing by employing deep-learning algorithms for feature extraction and real-time recognition.

THE RISE OF MULTIMODAL AI: A QUICK REVIEW OF GPT-4V AND GEMINI

2025-06-21T11:42:25+03:00

Multimodal artificial intelligence (AI) systems— interpreting, synthesizing and reasoning heterogeneously over text, images, audio and video—represent a transformational boundary in AI research and application today. Some notable achievements in this area are Open AI GPT-4V (Vision) and Google DeepMind’s Gemini 1.5, both exemplifying the current coups of cross-modal representation learning and generative reasoning. This paper remarks critically and succinctly on these two flagship models, studying their architecture, modality fusion, functionality, and performance metrics. Emphasis is placed upon their performance towards visual question answering, multimodal dialogue, instruction following, and other tasks that are reasoning integrated because intelligence and perception working in harmony are needed. Moreover, we examine GPT-4V and Gemini 1.5 from the lenses of model size, scaling, fine-tuning, alignment, and generalization in downstream tasks. The debate looks at the major outstanding issues of multimodal AI: hallucinations, no interpretability, high computational cost, and others which remain the most important barriers to wider use and trust. Finally, we study the far-reaching effects

EXPERIMENTAL INVESTIGATION ON WORKABILITY, MECHANICAL PROPERTIES & THERMAL EFFECT OF POLYMER FIBER REINFORCED CONCRETE (STYRENE BUTADIENE RUBBER (SBR), POLYPROPYLENE ROPE) USING BRICK AND NATURAL AGGREGATE

2025-06-21T13:30:32+03:00

This study investigates the workability, mechanical properties, and thermal effects of polymer fiber-reinforced M20-grade concrete with varying aggregate compositions The experimental program evaluates four concrete mixes: (1) Control mix with 100% natural aggregate (without polymer and fiber), (2) 100% natural aggregate (with polymer and fiber), (3) 50% natural + 50% brick aggregate with polymer and fiber, (4) 100% brick aggregate with polymer and fiber, and assessment of thermal conductivity effect for each mix. A fixed 10% polymer ratio was maintained, while the polypropylene rope fiber content varied from 0% to 1%. The fresh concrete was evaluated for workability using the slump test, while hardened properties were assessed through compressive strength, split tensile strength, and flexural strength tests at 7 and 28 days also analyzing thermal conductivity and insulation performance. Results indicated that increasing fiber content reduced workability but improved mechanical properties up to an optimum fiber dosage. Incorporating brick aggregates enhanced thermal insulation properties but led to a minor reduction in mechanical strength. This study highlights the feasibility of polymer fiber-reinforced concrete with hybrid aggregates for sustainable and durable construction applications.