Volume 14 | Issue 2 |

 Abstract

The need of facial skincare is growing and the market is huge. Skin kinds, skin problems and skin tones can all be used to classify facial skin. For customized skin care, an accurate skin type classification is essential. This paper investigates the classification of facial skin types using Convolutional Neural Networks (CNNs) and makes recommendations for how CNNs can be used to efficiently categorize skin types and provide a useful tool for individualized dermatology and skincare. This study opens the door for further developments in individualized dermatological care by adding deep learning technology into the skincare sector.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Facial Skin Type; Convolutional Neural Network; Dermatology; Deep Learning

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

The Internet of Things (IoT) represents a revolutionary shift in the way devices, systems, and people are interconnected, enabling automation and efficiency across various domains, from healthcare to smart cities. However, the massive scale and complexity of IoT networks introduce significant security challenges. In this paper, we explore various security patterns and architectural frameworks designed to address these challenges. The goal is to provide a comprehensive overview of the current state of IoT security, identify critical vulnerabilities, and discuss evolving patterns and architectures that strengthen IoT systems against emerging threats.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Exploring Patterns and Architectures for Strengthening IoT Security

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Modern advancements in technology, such as sensors, satellites, and predictive models, have significantly improved early warning systems for disasters like hurricanes, floods, and tsunamis. These tools enable experts to issue timely alerts and evacuate at-risk populations. The Internet of Things (IoT) plays a crucial role in disaster response by facilitating real-time monitoring of environmental conditions and infrastructure integrity. IoT-driven solutions enhance damage assessment, monitor vulnerable communities, provide real-time data, and predict disaster impacts. By integrating IoT into disaster management frameworks, response times can be improved, coordination between agencies can be strengthened, and recovery efforts can be accelerated. This paper explores IoT-based disaster recovery, highlighting key enabling technologies and proposing an innovative algorithm for establishing temporary network connections in disaster-affected areas. [1].

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Analytics, Coordination, Disaster Response, IOT, Predictive, Sensor, Social Media, Resilient

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Smart cities are rapidly emerging as the backbone of modern urban development, driven by the integration of cutting-edge technologies such as the Internet of Things (IoT), artificial intelligence (AI), and big data. These technologies enhance operational efficiency, improve public services, and contribute to sustainable development. However, they also introduce significant cybersecurity challenges, as the interconnected nature of smart city infrastructure creates vulnerabilities that can be exploited by malicious actors. From transportation systems and energy grids to healthcare and public safety networks, critical urban infrastructure faces a growing risk of cyberattacks that can disrupt essential services, compromise public safety, and lead to significant financial losses. This paper explores the cybersecurity risks facing smart cities, with an emphasis on safeguarding critical infrastructure through a multi-layered defense strategy. It presents a comprehensive analysis of the most common threats, including ransomware attacks, Distributed Denial of Service (DDoS), data breaches, and unauthorized access to IoT devices. The paper proposes a multi-layered approach to security, encompassing network protection, application security, endpoint hardening, data encryption, and real-time monitoring. Each layer serves as a critical component of a holistic defense model, aiming to mitigate risks and enhance resilience against cyber threats. Furthermore, this research highlights the importance of collaboration between public and private sectors, smart city administrators, and cybersecurity professionals. It discusses the necessity of developing proactive incident response plans and implementing advanced technologies such as AI-powered threat detection and blockchain for secure data management. By adopting a multi-layered security strategy, cities can better protect their urban infrastructure and ensure the continued safe operation of smart technologies.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Smart Cities, Cybersecurity, Critical Infrastructure, IoT, Multi-Layered Security, Threat Mitigation, Incident Response

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

The process of compression becomes vital to maximize data storage and transmission optimization and especially applies to big data environments. Digital data continues to expand rapidly thus demanding more effective compression methods to become essential. Many types of complex datasets containing high numbers of dimensions prove incompatible with standard compression methods like Huffman coding and Run-Length Encoding. Generative AI models represent new methods for compression, including Variational Autoencoders (VAEs), Generative Adversarial Networks (GANs), and Vector Quantized Auto encoders (VQ-VAEs), which are emerged as promising alternatives. This study analyzes how generative AI systems develop image and video compression process output and checks their output quality along with resource usage metrics. The main function of compression techniques is to decrease redundant data without compromising crucial information that needs for efficient storage and transmission. The advancement of generative models now enables them to acquire complicated data distributions along with compressed representations which maintain essential aspects from original data. The core purpose of this investigation is to analyze different generative AI models for data compression based on their effectiveness and evaluate their strengths and weaknesses when processing images and video content. In addition the study will identify optimal models for particular applications through performance-based assessments that measure compression effectiveness and reconstructive quality and computational efficiency. Some generative models demonstrate their competence by reaching high compression ratios alongside preserving quality standards in lossy compression operations. These models either present realistic visual reconstructions although they demand increased computational resources. This research into generative model comparison with traditional methods reveals significant findings suitable for various application-related decisions making

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Generative AI, Data Compression, Compression Ratio, Reconstruction Quality, Computational Efficiency

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

The proposed research is based on a smart strategy for encrypting audio files using Chaotic map and DNA Encoding. It targets the important parts of sound to protect from unethical use. Chaotic map puzzles hackers and DNA adds a layer of mystery, making it very secure indeed. Algorithm is tested for various statistical test like correlation analysis, information entropy analysis, scatter plots diagrams, histogram, mean squared error, PSNR and NIST tests, these show it's strong and dependable. A great thing is, once you decode the audio, it sounds just as good as before. So, if you're into sharing secret tunes, this approach suits you well.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Sound scrambling, DNA encoding, Multimedia data, Henon map, Hybrid chaotic map

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Road accidents remain a major global challenge, requiring rapid and effective emergency response mechanisms. Traditional accident detection methods rely on manual reports or eyewitness accounts, resulting in delays in response times and increased severity of the consequences. This paper presents a CNN-based accident detection and emergency response system using YOLOv8, an advanced deep learning algorithm for real-time video analysis. The system processes live video surveillance images to detect accidents quickly, records critical details such as location, time, and severity, and triggers immediate alerts to the emergency services. The proposed system automates accident monitoring, integrating machine learning and computer vision technologies to achieve accuracy while minimizing false positives and negatives. It features a user-friendly interface for traffic authorities and emergency responders, a robust detection module powered by convolutional neural networks, and a scalable database for efficient storage and analysis of accident data. The methodology includes data collection, pre-processing, model training, evaluation, and deployment, guaranteeing the adaptability of the system, deployment, and ensuring adaptability in a variety of environments, including urban areas, freeways, and high-traffic areas. The preliminary results from controlled experiments demonstrate the system's potential to significantly reduce emergency response times and coordination between emergency services. By filling the gaps left by traditional detection systems, this study provides a scalable and reliable solution that contributes to intelligent transportation systems and improves public safety through real-time monitoring notifications and effective emergency management.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Accident Detection; Convolutional Neural Networks; Emergency Response; YOLOv8

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Nowadays, The Internet of Things (IoT) has made our lives more reliable and efficient in multiple ways. IoT is a rapidly emerging technology in the consumer, business, industrial, and social ecosystems. IoT networks use the communication technologies, such as IoT devices, to share and spread information applications and hardware. As such, Distributed Denial of Service attacks use multiple connected devices executed by botnets and creates many harmful and dangerous threats to the security of IoT networks. Attackers can analyze and attack IoT devices as part of botnets to launch DDoS attacks by taking advantage of their flaws and targeting the server by sending a flood of messages and creating internet traffic then the system halts and reduces the performance of the system. In this research, when an attacker sends a flood of fraud messages, then some alert notifications, warning messages, and alarms are triggered in the victim's machine to avoid data loss. Also includes secondary data (research papers, case studies, and past cybercrime studies) to detect the threats and prevent them in the network. This presenting paper throws light on the prevention and techniques of DDoS attacks in IoT.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Internet of Things (IoT), Alert notification, Internet traffic, Distributed Denial of Service (DDoS) attack, Botnet.

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Ant Colony Optimization (ACO) is a bio-inspired optimization technique based on the foraging behavior of ants. It has been widely applied to complex problems in various domains, including logistics, robotics, and artificial intelligence. This paper provides an overview of ACO, highlighting its principles, applications, and various adaptations. It also discusses the challenges and opportunities for future research in ACO, with a focus on hybrid approaches and emerging technologies. The goal is to shed light on the evolving role of ACO in solving real-world optimization problems.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Ant Colony Optimization (ACO), Swarm Intelligence, Ad-hoc Network, TSP, VRP

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Social media sites like Twitter have become increasingly important over the past few years in creating awareness and provoking debate around social issues like gender-based violence. Social media sites offer extensive user-generated content reflecting public sentiment and engagement on specialised subjects. This study uses clustering, community detection, and sentiment analysis to analyse public debate around gender violence on Twitter. The research aims at a corpus of 6,799 tweets representing varying gender violence-related discourse. The corpus was initially preprocessed to remove extraneous content such as stopwords, URLs, and uncorrelated tweets. The tweets were then vectorised using TF-IDF (Term Frequency-Inverse Document Frequency) to identify the meaningful attributes. K-Means clustering was employed to group similar tweets, while Louvain's community detection algorithm was employed to identify individual communities of users discussing gender violence. Sentiment analysis was done to classify tweets as positive, negative, or neutral. At the same time, Different evaluation measures, such as Modularity, Silhouette Score, and Davies-Bouldin Index, were used to analyse the efficiency of clustering and community detection. The objective of this study is to create a machine-learning model that classifies gender-related tweets into one of five categories: sexual violence, emotional violence, harmful cultural practices, physical violence, and economic violence. The research indicates that most problems are unreported, and information is filtered through perpetuating mechanisms and consequences for Indian society. The mental well-being of the women and children also impacts the problems in question. With the support of increasing efforts to promote women's empowerment and mainstreaming gender equality, we can create social awareness on social networking sites with actions that favour women's and girls' development and livelihoods.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Gender violence, Women empowerment, Gender equality, Social media activism, Data analysis

  License

Creative Commons Attribution 4.0 and The Open Definition