Missing Person Identification in Crowds

Overview

A face-recognition surveillance and communication system for locating missing persons at Simhastha Ujjain, a mass gathering attended by millions of people. It automates identification of missing persons from live video feeds, notifies police and families when a match is found, and gives both the public and police staff a simple interface for filing reports and monitoring alerts.

A frontend deployment is available at missing-person-identification.netlify.app; the backend deployment was removed due to insufficient Azure credits.

What it does

• Real-time face recognition: identifies missing persons from live video feeds, speeding up detection.
• Live video monitoring: continuously compares detected faces against stored records for potential matches.
• Image enhancement: GFPGAN improves the quality of user-uploaded photos to support accurate identification.
• Instant communication: Twilio sends notifications to police and families when a match is detected.
• Centralized dashboard: police can access missing-person reports, live feeds and real-time alerts in one place.
• Scalable architecture: designed to handle the volume of data generated at an event the size of Simhastha Ujjain.
• User-friendly interfaces: separate, accessible interfaces for the public (filing reports) and police staff (monitoring alerts).

How it's built

Technologies used, as listed in the README:

• Backend / frontend: Django, React JS, MySQL
• Vision: OpenCV, Google Cloud Vision API, Hugging Face's GFPGAN
• Messaging: Twilio
• Cloud and infrastructure: Google Cloud, Azure Blob Storage, Redis

Context

Built by team "ERROR 404 : CHANGE FOUND?" — six members, led by me — for the Simhastha Ujjain mass-gathering problem statement listed in the README, at Smart India Hackathon 2024, where it placed in the top 45.