CoolFace
Apppublic

MLengineers/distracted-driver-detection

sourceHugging Facemitupdated 8mo agoView on Hugging Face
0likes
App README

🚗 Distracted Driver Detection

A real-time distracted driver detection system using deep learning with 78.33% accuracy.

Features

  • —📷 Image Upload - Analyze driver photos
  • —📹 Live Webcam - Real-time detection
  • —🎯 10 Classes - Detects various distractions

Detected Behaviors

ClassBehavior
c0Safe driving ✅
c1Texting - right hand
c2Phone call - right hand
c3Texting - left hand
c4Phone call - left hand
c5Operating radio
c6Drinking
c7Reaching behind
c8Hair and makeup
c9Talking to passenger

Model Details

  • —Architecture: Custom CNN (27M parameters)
  • —Training: 15 epochs on 22,424 images
  • —Accuracy: 78.33% validation accuracy
  • —Input Size: 224×224 RGB images

Usage

  1. 1.Click "Upload Image" or "Live Webcam"
  2. 2.Provide a driver image
  3. 3.View detection results with confidence scores

Tech Stack

  • —TensorFlow/Keras
  • —Flask
  • —OpenCV
  • —Modern CSS UI

Built with ❤️ using Deep Learning