AI-Powered Geo-Intelligent Lost & Found System for Smart Campus Environments

Losing personal items on university campuses is a recurring challenge that often leads to low recovery rates due to fragmented and manual retrieval processes. This paper introduces "Lost & Found Mycampus," a digital system that uses AI and mapping tools (GIS) to make finding lost things much easier. Unlike traditional logbooks, our system employs a pre-trained ResNet18 Convolutional Neural Network (CNN) for automated high-precision visual feature extraction, coupled with an interactive geospatial reporting interface. We formulated a matching engine that calculates a "Total Score" using a weighted combination of visual similarity, spatial proximity (via the Haversine distance formula), and temporal decay. Specifically, we used a standard ResNet18 model to automatically pull out a vector of 512 distinct matching features from item photos, which we then cross-reference with where the item was last seen. Our experimental observations indicate that our visual matching approach achieves rapid classification cycles under 500ms and significantly optimizes the discovery-to-return lifecycle. The results demonstrate that the proposed method achieves an 88.6% recovery rate, outperforming traditional manual baselines by a significant margin. This framework provides a scalable, low-latency solution for modern "Smart Campus" adoption, effectively bridging the gap between physical loss and digital recovery.