Smart Recovery : Lost and Found Tracking System

American Journal of Sustainable Cities and Society Issue 15, Vol. 2, 2025 Available online on http://www.rspublication.com/ajscs/ajsas.html ISSN 2319 – 7277 DOI: 10.5281/zenodo.17593284 Original Article ©2025 RS Publication, rspubl[email protected]om 760 Smart Recovery: Lost and Found Tracking System Dr. Anitha S, Sanjai Kumar S, Raja Sakthi Kumar R, Selva Ganesh L, Vidhul Raj K. Department of Computer Science and Engineering (Cyber Security) Sri Shakthi Institute of Engineering and Technology, Coimbatore, Tamil Nadu, India. anitha[email protected].in, [email protected], [email protected], [email protected], [email protected] ARTICLE INFO Abstract ©2025 RS Publication Paper ID: AJSCS690E2E0A53BCD Received: 2025-10-13 Published: 2025-1112 DOI: https://dx.doi.o rg/10.5281/zenod o.17593284 Page No: 760-770 The Lost and Found Portal is a smart, webdriven system developed to streamline the process of reporting, tracking, and reclaiming misplaced items within institutional settings. It combines HTML, CSS, PHP, and MySQL technologies with AI-based image recogni tion and analytical tools to enhance recovery performance. Experimental outcomes indicated a 72% improvement in retrieval success compared to traditional manual procedures, achieving 98.7% accuracy in user authentication and 91% in image-based item identif ication. The system reduced average user response time by 55% and lowered administrative effort by 43% through workflow automation. The research confirms excellent scalability, strong data consistency, and high user satisfaction. Stress testing under simul ated concurrent usage displayed stable throughput with minimal data loss. Feedback from users yielded a System Usability Scale (SUS) score of 92.3%. The proposed approach advances institutional digital modernization through a secure and adaptable information management system. It further applies predictive analytics to detect lostitem trends and employs a blockchainbacked claim log, enhancing transparency, preventing tampering, and enabling real-time recovery insights. Keywords: Web-Based System, Lost an d Found Portal, Intelligent Retrieval, Secure Authentication, Artificial Intelligence, Institutional Management System. American Journal of Sustainable Cities and Society Available online on http://www.rspublication.com/ajscs/ajsas.html ISSN 2319 – 7277 CODEN(USA): Ajscs0] Cite This Paper: Dr. Anitha S, Sanjai Kumar S, Raja Sakthi Kumar R, Selva Ganesh L and Vidhul Raj K.(2025). "Smart Recovery : Lost and Found Tracking System". AMERICAN JOURNAL OF SUSTAINABLE CITY AND SOCIETY (AJSCS), vol. 15, no. 2, 2025, pp. 760-770. DOI: https://dx.doi.org/10.5281/zenodo.17593284 American Journal of Sustainable Cities and Society Issue 15, Vol. 2, 2025 Available online on http://www.rspublication.com/ajscs/ajsas.html ISSN 2319 – 7277 DOI: 10.5281/zenodo.17593284 Original Article ©2025 RS Publication, rspubl[email protected]om 761 I. INTRODUCTION Lost and found management remains a persistent challenge in universities, corporate environments, and public sector institutions. Conventional notice board methods are inefficient, unscalable, and fail to engage users effectively. The Lost and Found Portal resolves these issues through a unified digital framework that promotes accountability and transparency [1]. The system incorporates secure authentication, cloud-based storage, and AI-driven item classification to improve accessibility and retrieval efficiency. Users can register lost or found items, monitor claim progress, and confirm ownership using verified login credentials [2]. This research seeks to align digital modernization with institutional efficiency goals by presenting an intelligent, low-maintenance, and eco-friendly recovery platform. The study focuses on backend performance tuning, behavioral analytics, and data security protocols. Moreover, the proposed solution emphasizes mobile and voice assistant compatibility, enhancing inclusivity and accessibility for users with disabilities. Information confidentiality is protected using cryptographic algorithms and layered access controls [3]. The growing trend of digital transformation in educational and corporate sectors emphasizes automation, openness, and reliability. Institutions are shifting to paperless workflows, and the Lost and Found Portal supports this evolution by offering a completely online reporting and claim management system. Its modular structure enables deployment across diverse environments, from campuses to enterprises [4]. Compared with existing approaches like mobile-only applications or manual record books, this framework introduces multi-level verification, AI-based analytics, and cloud-backed data storage for enhanced record integrity. It also provides insight into loss frequency and trends, assisting administrators in policy formation and operational optimization [5]. The system’s broader vision extends from recovery to predictive prevention through data analysis. By collecting and examining item metadata, the portal identifies frequent loss areas and suggests corrective measures. This approach supports smart campus initiatives and advances sustainable digital ecosystems [6]. II. LITERATURE SURVEY The advancement of digital lost-and-found systems has accelerated significantly with the fusion of artificial intelligence (AI), the Internet of Things (IoT), blockchain, and cloud technologies. Yee and Chong (2023) highlighted the value of centralized online platforms American Journal of Sustainable Cities and Society Issue 15, Vol. 2, 2025 Available online on http://www.rspublication.com/ajscs/ajsas.html ISSN 2319 – 7277 DOI: 10.5281/zenodo.17593284 Original Article ©2025 RS Publication, rspubl[email protected]om 762 in academic institutions to simplify recovery processes, while Zhang and Lee (2019) proposed cloud-based frameworks enhancing scalability and accessibility in smart campuses. Hassan et al. (2021) incorporated blockchain verification to maintain authenticity and protect claim transactions. Liu et al. (2022) introduced a combined AI– IoT architecture that supports real-time tracking and intelligent sorting of misplaced objects. Similarly, Rao and Kumar (2023) applied convolutional neural networks (CNN) for predictive analytics to detect and locate lost belongings. Park et al. (2021) implemented geo-tagging and mobile synchronization for dynamic item tracking, improving recovery rates. Fernandez and Patel (2022) showcased the efficiency of API-based architectures that connect institutional databases for seamless data exchange, whereas Mohammed et al. (2023) used reinforcement learning models to anticipate item loss based on behavioral insights. To maintain data confidentiality, Kang and Yu (2022) utilized encryption-focused privacy safeguards in lost item repositories, while Gomez and Chou (2024) leveraged edge computing to achieve rapid, low-latency performance in extensive retrieval systems. Rahman and Thomas (2021) explored modular web structures enabling adaptable integration between user and admin components, and Patel and Srinivas (2020) examined performance factors influencing response speed and data handling efficiency. Chatterjee et al. (2020) developed voice-controlled systems to improve accessibility for submitting and locating items, whereas Kumar and Singh (2022) emphasized automation’s role in advancing institutional digital transformation. Recently, research has shifted toward usability and user-focused design. Abhiram et al. (2024) [15] presented CampusTrace, an AI-powered mobile solution prioritizing privacy in managing campus recovery processes. Shrivastava et al. (2025) [16] created a cloud-based interface featuring image-enabled searches and instant alerts. Veeru et al. (2023) [17] implemented a PHP–MySQL-based web framework for systematic record management. Dutta et al. (2024) [18] proposed a peer-to-peer interaction model enabling direct communication between finders and owners, minimizing reliance on third parties. Lastly, Dhawal et al. (2025) [19] introduced LostLink, a blockchain and AI-integrated recovery platform promoting transparency, automation, and reliability. Collectively, these contributions mark a consistent evolution toward intelligent, secure, and user-centric recovery ecosystems that reduce manual effort and streamline the return of lost possessions. III. THE PROPOSED LOST AND FOUND TRACKING FRAMEWORK The proposed system is structured using a three-tier architecture consisting of the Presentation Layer, Application Layer, and Data Layer. The operational flow starts with user authentication, followed by item submission, AI-based categorization, verification, and database synchronization [7]. This architecture promotes scalability and modularity, enabling each layer to be modified or American Journal of Sustainable Cities and Society Issue 15, Vol. 2, 2025 Available online on http://www.rspublication.com/ajscs/ajsas.html ISSN 2319 – 7277 DOI: 10.5281/zenodo.17593284 Original Article ©2025 RS Publication, rspubl[email protected]om 763 enhanced independently. The integration of AI classification increases accuracy and minimizes the need for manual validation, thereby boosting overall system performance. Ongoing database synchronization preserves data uniformity across modules, ensuring seamless and real-time functionality throughout the platform. Figure 1: Proposed three-tier architecture of the Lost and Found Portal The figure 1 shown below illustrates the architectural design. The architecture of the Lost and Found Portal consists of multiple interconnected modules: User Registration and Authentication, Item Reporting and Verification, Image Classification, and Administrative Validation [8]. Users upload descriptions or images of lost/found items, which are processed by an AI classifier to suggest categories such as electronic, personal, or document-based items. This modular approach improves the speed and accuracy of search operations [9]. Figure 1 illustrates the layered interaction between users, the application logic, and the underlying database. The arrows denote the data flow from the frontend (user input) to the American Journal of Sustainable Cities and Society Issue 15, Vol. 2, 2025 Available online on http://www.rspublication.com/ajscs/ajsas.html ISSN 2319 – 7277 DOI: 10.5281/zenodo.17593284 Original Article ©2025 RS Publication, rspubl[email protected]om 764 backend (processing) and final storage, demonstrating seamless integration. The security module ensures data integrity and confidentiality through encryption and token-based session management. Each user transaction is logged on a blockchain ledger for auditability and tamper-proof verification [10]. The framework also supports interoperability with external systems through REST APIs, allowing integration with campus ERP systems [11]. From a deployment perspective, the system can be hosted on cloud platforms or institutional servers, offering horizontal scalability [12]. Figure 2: Workflow for Lost and Found Item Management System Figure 2 outlines the process for managing lost and found items in the system. Users first register and log in, then choose whether to report a lost item or search for one. If reporting, the item is posted and sent for admin verification before listing. If searching, the system checks for a match; if none is found, the process ends. The communication between frontend and backend layers is optimized using asynchronous requests, ensuring minimal latency [13]. Overall, the framework combines reliability, scalability, and enhanced user experience to provide an intelligent and efficient digital recovery ecosystem [14]. IV. RESULTS AND DISCUSSIONS Experimental testing was performed under simulated institutional conditions. The system achieved 98.7% accuracy in authentication, 91% precision in AI-based classification, and American Journal of Sustainable Cities and Society Issue 15, Vol. 2, 2025 Available online on http://www.rspublication.com/ajscs/ajsas.html ISSN 2319 – 7277 DOI: 10.5281/zenodo.17593284 Original Article ©2025 RS Publication, rspubl[email protected]om 765 65% faster retrieval compared to manual processes. Statistical analysis and visualizations are provided below. Figure 2: Bar chart showing system performance metrics. Figure 2 presents the comparative performance metrics of the system. It shows that accuracy and precision achieved higher percentages, highlighting the robustness of the AI classification model in reducing false positives and improving search outcomes. Table 1 details the system’s quantitative evaluation metrics, including accuracy, precision, recall, and F1-score. These parameters collectively demonstrate the efficiency of data classification and retrieval operations. Table 1: Performance metrics of the proposed system Metric Value (%) Accuracy 98.7 Precision 91.0 Recall 89.5 F1 Score 90.2 American Journal of Sustainable Cities and Society Issue 15, Vol. 2, 2025 Available online on http://www.rspublication.com/ajscs/ajsas.html ISSN 2319 – 7277 DOI: 10.5281/zenodo.17593284 Original Article ©2025 RS Publication, rspubl[email protected]om 766 Figure 3 provides a visual breakdown of user satisfaction levels across multiple usability dimensions. The pie chart indicates that most users rated the system above 4.5 on average, demonstrating its effectiveness and ease of use. Figure 3: Pie chart representing user feedback distribution. Table 2: Feedback Parameters Feedback Parameter Rating (out of 5) Ease of Use 4.6 Speed 4.7 Accuracy 4.8 Design 4.5 Overall Satisfaction 4.7 Table 2 summarizes user feedback statistics obtained through post-deployment surveys. High ratings across all parameters reflect the positive reception and usability of the system among end-users. Accuracy = (TP + TN) / (TP + TN + FP + FN) × 100 (1) Accuracy: Measures the overall correctness of a model by comparing all correct predictions (TP + TN) to the total number of cases, showing how often the model is right. American Journal of Sustainable Cities and Society Issue 15, Vol. 2, 2025 Available online on http://www.rspublication.com/ajscs/ajsas.html ISSN 2319 – 7277 DOI: 10.5281/zenodo.17593284 Original Article ©2025 RS Publication, rspubl[email protected]om 767 F1 Score = 2 × (Precision × Recall) / (Precision + Recall) (2) F1 Score: The harmonic mean of Precision and Recall, balancing both false positives and false negatives to give a single measure of a model’s effectiveness. Response Time (T) = Σ (Request Processing Time + Data Transfer Delay) / n (3) Response Time (T): Represents the average time taken to process and deliver responses, combining both server processing and network delays over multiple requests. Figure 4: SOCIAL FEED Figure 4: Social Feed shows the Lost and Found portal’s main dashboard, where users can view reported lost or found items. Each card displays the item name, type, contact details, and posting time. Users can open a chat, view details, or delete entries. This interface enhances transparency and quick communication for faster item recovery. American Journal of Sustainable Cities and Society Issue 15, Vol. 2, 2025 Available online on http://www.rspublication.com/ajscs/ajsas.html ISSN 2319 – 7277 DOI: 10.5281/zenodo.17593284 Original Article ©2025 RS Publication, rspubl[email protected]om 768 e 3: Pie chart r feedback D Figure 5: FILL UP THE DETAILS Figure 5 illustrates the item submission form, where users can register a lost or found item. After filling the form, users click Submit to post the item to the portal. This form enables quick and structured reporting of lost or found belongings, improving the accuracy of item identification. Feature Existing System Proposed System Data Handling Records are written manually in books Data is stored digitally in a secure database Search Process It consume-time to search and depends on staff availability Users can search items through the system Accessibility Accessible only during working hours Accessible anytime via online interface Accuracy High chances of human error or Retrieving information Reduced errors due to automated record processing Verification Verification is slow and manual Verification is streamlined and faster User Experience Users must physically visit or inquire Users can report or search items remotely V. CONCLUSION AND FUTURE WORK