Melanoma is a highly aggressive form of skin cancer where early detection significantly improves survival. Automated segmentation assists dermatologists by providing accurate lesion boundaries, but traditional approaches struggle with noise, illumination variations, and low contrast. Building on challenges noted in prior work [1], this study proposes a modified ResUNet model trained on ISIC 2018, integrating enhanced feature extraction and boundary refinement. The model achieves an 85
The merger between Artificial Intelligence (AI) and the Internet of Things (IoT) is rapidly changing healthcare, smart homes, mobility, and infrastructure in the population. But such assimilation intensifies deep ethical and legal dangers. Privacy is reduced to symbolic consent because wearables constantly track sensitive types of biometric information [1]; equity is marginalized despite life-threatening discrimination like racial disparity in healthcare [7]; transparency is rhetorical, and accountability is diffused to many actors without discussing the harms [3]; and autonomy is eroded silently by automation in smart environments [2]. Although prior studies have examined the algorithmic behavior of IoT [4], wearable privacy [5], and AI fairness [6], the literature remains dominated by abstract principles. The paper presents a model, Ethico-Legal Accountability and Transparency in IoT (ELATI), that is an integrated framework that realizes ethics in an enforceable, cross-domain manner. In comparison with the previous practices, ELATI goes further to aspirational principles and entrenches ethics in technical, legal, and organizational practice. It incorporates technical protection (privacy-preserving analytics, real-time fairness checks, explainability dashboards), governance (adaptive regulation, distributed accountability, liability insurance), and user-friendly protections (autonomy protection, transparent consent, customizable nudges). ELATI uses practical case studies to demonstrate how it works: healthcare wearables counter-acting racial bias in pulse oximeters [7]; smart homes giving users a transparency dashboard and opt-out options [2]; and the interplay of IoT and social media needs cross-border responsibility on misinformation [8]. The originality of this research is in changing ethics as some abstract promises into the practical, enforcing, cross-sectoral, globally versatile, and practically enforceable mechanisms. ELATI offers a policy roadmap that policymakers, developers, and regulators can use to make AI-IoT systems respectful, equitable, trustworthy, and just and remain sensitive to varied social and legal settings.
With the rapid growth of vehicle ownership and the continuous expansion of urban areas, traditional parking practices are becoming increasingly inadequate. These outdated systems often lead to long waiting lines, traffic congestion, and unnecessary fuel consumption as drivers circle around in search of available spaces. To address these issues, many cities and researchers have begun turning to IoT-based smart parking solutions, which offer features such as automated payment processing, real-time monitoring of vacant slots, and an overall more convenient experience for users. Recent literature highlights a wide range of approaches, including the use of different sensor technologies, microcontrollers, machine learning techniques, and mobile application interfaces. Several studies also examine system architecture, the integration of hardware and software components, and the usability of the platforms from an end-user perspective. Collectively, these works demonstrate the potential of smart parking systems to ease congestion, reduce fuel wastage, and contribute to more sustainable urban mobility. The purpose of this paper is to consolidate these developments, provide a clear picture of the present landscape of IoT-enabled parking systems, and outline possible future directions that could support smarter and more efficient urban environments.
Protecting the privacy of data stored with third-party providers has become a critical concern, driving the need for privacy-preserving solutions such as searchable encryption (SE). Despite these protections, such systems are vulnerable to file injection attacks, where adversaries exploit indexing mechanisms to compromise data integrity and gain insights of users’ search queries. Those attacks represent a significant breach of data confidentiality and weaken the reliability of cloud-based outsourced systems. In this work, a secure phrase-based searchable encryption approach is proposed to mitigate file injection attacks. This approach adopts SHA-256 hashing to securely determine bloom filter positions for efficient keyword searches, and uses AES encryption in Cipher Block Chaining (CBC) mode to protect both the index and query keywords. Experimental results demonstrate that the proposed approach achieves a search accuracy of 97.6
In contemporary society, many individuals encounter mobility challenges resulting from ailments such as spinal cord injuries, advanced arthritis, neurological conditions including multiple sclerosis and cerebral palsy, or limb amputations. These ailments effect their autonomy and capacity to engage in routine activities, making wheelchairs essential assistive instruments. However traditional wheelchairs often fail to meet the requirements of users with severe limb or joint impairments. The paper presents a dual function wheelchair system that uses Electromyographic and accelerometric techniques to facilitate intuitive navigation. The system uses MPU6050 accelerometer to detect head movements, or Electromyography based navigation based navigational commands. To make the wheelchair practically useable for detecting and navigating through the obstacles, Ultrasonic sensors are incorporated. This innovative design enhances user autonomy, supports varying levels of mobility improving the overall functionality of assistive mobility devices.
