🚗 Vehicle & People Surveillance CV System

A multi-stage computer vision pipeline that detects cars and people, then zooms in to identify license plates and faces.

🔍 Demo

🏗️ Architecture

Input Frame → [Primary Detector] → Car ROI → [Plate Detector] → Plate Crop → Person ROI → [Face Detector] → Face Crop

📊 Model Performance (10 epochs)

Model	mAP50	mAP50-95
Primary (Car+Person)	0.807	0.479
License Plate	0.83	0.539
Face Detector	0.864	0.57

🗂️ Datasets

• Car Object Detection (Kaggle)
• People Detection (Kaggle)
• Indonesian Vehicle License Plate (Kaggle)
• Face Detection Dataset (Kaggle)

🚀 Run It Yourself

🛠️ Tech Stack

• YOLOv8 (Ultralytics)
• OpenCV
• PyTorch
• Python 3.12