Real-time Object Detection

Overview

Detects, labels and counts objects in a live camera feed with confidence scores. Python backend, React frontend, and a GitHub → AWS CI/CD pipeline. Open source.

The problem

Object detection demos usually stop at a notebook. This one is a complete app: model, API, UI and deployment pipeline.

How it was built

A custom-trained YOLOv5 model served through a Python API so any frontend can consume it, a React client that shows detections over the live feed, and a GitHub → AWS CI/CD pipeline for deployments.

What it does

• 01Instant detection and labelling in live video
• 02Counts detected objects with an average confidence score
• 03Model behind an API, reusable across clients

What I took from it

Natural extensions: CCTV anomaly detection, collision avoidance in vehicles, retail analytics, wildlife monitoring.