In the era of artificial intelligence, the proliferation of AI-generated images has blurred the boundaries between reality and digital fabrication. Technologies such as Generative Adversarial Networks (GANs) have enabled the creation of highly realistic synthetic images—commonly known as deepfakes—which pose substantial challenges in domains like digital media, cybersecurity, and legal forensics. While these advancements offer innovative applications in entertainment and simulation, their potential misuse can lead to misinformation, identity theft, and erosion of public trust. This project proposes an AI-powered image authenticity detection system that leverages a Convolutional Neural Network (CNN) to accurately classify images as either real or AI-generated. The system is built with an intuitive graphical user interface (GUI) that allows users to upload and analyse images in both individual and batch modes. Key features include real-time prediction with confidence scoring, visual result displays, confusion matrix generation, and performance metrics such as accuracy, precision, and recall. The model achieves an overall classification accuracy of 82.7%, demonstrating strong potential for real-world applications in detecting synthetic media. By combining deep learning techniques with user-centric design, the system provides a practical and transparent solution for addressing the rising concerns of digital image manipulation. It serves as a critical tool for enhancing media authenticity and combating the spread of AI-generated misinformation.

This study examines the suitability of piezoelectric floor mat systems for harvesting energy in high traffic areas like the student centers. The study is aimed at solving the problems of small energy production, toxic materials and the ability to scale up current piezoelectric energy harvesting systems. The study involves experimental simulation of using 40 piezo transducers, a 2W02G rectifier, two 2F, 5.5V super capacitors for energy storage, an ND0603PC booster amplifier for output regulation and two LiPo batteries in series, to supply stable power to a case study Centre. Both the supporting circuit diagram and MATLAB/Simulink simulation were utilized to show that this system works well for independent power generation. Simulations and tests on circuits reveal that the system delivers an average output power greater than the required standard, 400–600 μW per step versus 134.2 μW per step. Rectifying the energy from 1,000 steps yields AC voltages varying from 20–80V which are then changed to DC at 18–75V. At the beginning, the super capacitors charge with 5–6V to last for 10–30 seconds before leveling off at 3.7–5.5V and the LiPo batteries provide about 5–20 mAh after being active for 10 minutes. Trials show that the device produces constant electricity under various stress tests, showing good conversion, storage and release of energy for powering small electronic devices. The results confirm that piezoelectric floor mats can be used affordably to produce energy anywhere in busy areas, thereby aiding efforts to make urban environments and the planet more sustainable. In the future, more experiments and improvements are required for deploying the technology on a wider scale.

This project presents the development of AI-based facial recognition smart glasses designed to assist visually challenged individuals in identifying people around them. The smart glasses integrate a compact camera with an AI-powered facial recognition system to detect and recognize faces in real time. The recognized faces are then conveyed to the user via an audio output system, enabling seamless interaction in social environments. The system utilizes machine learning algorithms to enhance accuracy and adaptability, allowing users to register and recall known faces. The proposed solution aims to improve the independence and confidence of visually impaired individuals by providing an accessible and user-friendly assistive technology. Through rigorous testing and optimization, the smart glasses demonstrate significant potential in enhancing the daily lives of visually challenged users.

The exponential growth of academic literature presents significant challenges for researchers in conducting comprehensive literature reviews and maintaining current knowledge in their fields. Traditional research methodologies often prove inadequate for processing the vast volumes of information available across multiple databases and repositories (Chen et al., 2024; Rodriguez & Kim, 2023). This study introduces a novel agentic artificial intelligence framework designed to enhance research productivity through intelligent automation of literature discovery and report generation processes. The proposed system employs a dual-agent architecture comprising a specialized Search Agent responsible for multi-database literature discovery and source quality assessment, and a Drafting Agent focused on content analysis, synthesis, and coherent report generation (Thompson & Williams, 2024). Through empirical evaluation involving 150 research tasks across 15 academic domains, our framework demonstrated substantial improvements over traditional research methods: 55% reduction in time requirements (from 18.7 to 8.3 days average), 23% improvement in source coverage (from 77% to 100%), 60% reduction in cost per literature review (from ,847 to ,139), and 28% increase in user satisfaction scores (from 3.2 to 4.1 out of 5.0). The system maintains high quality standards with an average quality score of 4.2/5.0 compared to 3.9/5.0 for traditional methods (Anderson et al., 2024). Domain-specific analysis reveals varying effectiveness, with interdisciplinary research showing the highest performance gains (68% time savings, 91% user satisfaction), followed by STEM disciplines (62% time savings, 94% satisfaction). The framework addresses critical challenges in academic research including information overload, source verification, and synthesis complexity while maintaining scholarly rigor and citation accuracy (Martinez & Lee, 2023). Implementation results demonstrate the practical viability of agentic AI systems in academic research contexts, providing a scalable solution for institutions seeking to enhance research productivity and quality.

This paper examines how linear regression in machine learning enables the prediction of air-ground path loss through environmental parameters such as temperature, humidity, and atmospheric pressure measurements. The paper demonstrates that temperature plays the most significant role in determining path loss, while humidity and atmospheric pressure contribute at a lower level. A high level of accuracy defines the linear regression model, which demonstrated efficient path loss prediction through a Mean Absolute Error (MAE) of 0.2995. The model demonstrates effective capabilities for system improvements during changing atmospheric conditions because the trend line shows the smooth progression of predicted and actual values. A hybrid model produced enhanced prediction accuracy when particle swarm optimization and gradient-boosting regressor parameters were optimized to establish the new model system. The optimized model substantially declined MAE to 0.0435, which verified its improved predictive capacity regarding absolute path loss values. A performance-maximized model resulted from tuning relevant parameters to set n_estimators equal to 56, learning rate to 0.1, and max_depth to 9. The optimized model accurately predicts path loss in communication networks, preparing it for on-site deployment. This research serves as a basis for further investigation, integrating other environmental elements, including wind speed, rainfall and elevation levels, and testing alternative state-of-the-art machine learning methods. Future improvements in these procedures can boost the flexibility and reliability of networks with an emphasis on air-ground systems. Research findings indicate that PSO-GBR hybrid models possess a high potential for path loss prediction, creating new possibilities for future air-ground communication systems and emerging technologies such as low-altitude satellites, air taxis, and unmanned aerial vehicles (UAVs).

This paper uses the Differential Evolution (DE) algorithm to optimize essential antenna parameters for maximum communication system performance. The research concentrated on improving antenna performance through performance indicator enhancement, including energy concentration and gain, while optimizing return loss, beamwidth, and efficiency to support reliable distant communications, especially under difficult operating conditions. The optimized parameters of the VHF air-to-ground antenna system reached 76.73 MHz frequency along with 45.00 dB gain, 10.00 degrees beamwidth and 0.92 efficiency, which demonstrates broad operational coverage while preserving low power dissipation. The impedance matching is effective because a return loss measurement shows 5.00 dB. The investigation applied the optimization structure to optimize a 5G millimetre-wave (mmWave) antenna system for dealing with propagation issues caused by high frequencies. The system achieved optimized parameters at 33.00 GHz frequency with 29.61 dB gain and 5.00 degrees beamwidth while maintaining 0.98 efficiency, proving its ability to handle dense deployments through minimal interference mechanisms and maximum spatial utilization capacity. The antenna's high return loss value of 28.14 dB demonstrates the signal integrity performance. A DE algorithm optimization succeeded with a validity cost of 1.173050, which confirmed its precision in calculations and demonstrated a steady MAE reduction, which proved the algorithm had reached its correct solution path. Findings from the research show that the DE algorithm optimizes antenna design problems in VHF frequencies and future 5G systems with both efficiency and robustness. The developed research findings deliver practical and methodological contributions to antenna optimization, enhancing energy efficiency and system performance for upcoming wireless communication networks.