CBIR-Based System for Locating Missing People and Objects

This project aims to develop an intelligent search system that helps locate missing individuals and lost belongings using image-based matching rather than text-based searches. The system utilizes a Convolutional Neural Network (CNN) for feature extraction, where two deep learning models (VGG and ResNet) were implemented and compared.

By converting each uploaded image into a vector of numerical features, the system is able to measure the similarity between images and retrieve the closest matches from a database using Content-Based Image Retrieval (CBIR) techniques. The evaluation showed that ResNet provided higher accuracy and better feature representation than VGG, resulting in more reliable matching results.

This solution can be applied by security agencies, public service platforms, and emergency response units to support real-time identification and reduce search time, helping families and communities in locating missing individuals more effectively.