AI Face Recognition & ReID Surveillance Platform by Jatin Kumar NagarAI Face Recognition & ReID Surveillance Platform by Jatin Kumar Nagar

AI Face Recognition & ReID Surveillance Platform

Portfolio project
Jatin Kumar Nagar
An AI-powered surveillance and identity platform that detects, recognises and tracks people across multiple camera feeds in real time, with forensic search over historical sightings and instant alerts for unknown faces.
Key features:
- Mission Control: multi-camera surveillance grid with local and remote (WebRTC) camera streams
- Real-time face detection and recognition (InsightFace) using 512-dimensional face embeddings
- Person re-identification (ReID) with body-based tracking, so identities persist even when faces are hidden
- Holistic perception: face mesh, 21-point hand tracking, 33-point pose landmarks and YOLOv8 object detection
- Neural Forensics: upload a photo and search every historical sighting by face similarity
- Identity registry with watchlists, restricted zones and automated WhatsApp alerts for unknown faces
- Analytics dashboard, operational log and PDF/CSV reports with age, gender and emotion insights
Architecture: three microservices: a React 19 frontend, a Node.js/Express orchestration server and a Python FastAPI inference worker, deployable with Docker Compose.
Tech stack: Python, FastAPI, InsightFace, MediaPipe, YOLOv8, OSNet, Node.js, Express, MongoDB, Socket.io, React 19, Redux Toolkit, Framer Motion, Docker
My role: designed and built the full system end to end, covering the AI inference pipeline, backend APIs, real-time streaming and the dashboard.

0 likes · 2 views · Nov 15, 2026

Built end to end: an AI surveillance platform with real-time face recognition, person re-identification, pose/hand/object detection, forensic face search and Wh

Tags