LifeLens: AI Panic Gesture Detection

• 

  Gesture detection

This project was born out of a strong desire to enhance safety for deaf and mute individuals, especially in emergency situations where verbal communication is not possible. We wanted to create an AI-powered system that could understand hand gestures as signals for help, making it easier to trigger emergency responses quickly and silently.

What I Learned: Throughout the development process, we explored and learned:

Real-time gesture detection using MediaPipe and OpenCV Building and training AI models using TensorFlow Integrating live location tracking and alert mechanisms Handling edge cases like varying lighting conditions and different backgrounds Enhancing UI/UX for accessibility and user-friendliness

How I Built It: Frontend: HTML, CSS, JavaScript with a unique 3D interactive interface

Backend: Python with AI-powered gesture recognition

Technologies Used: OpenCV, MediaPipe, TensorFlow, Geopy, smtplib, playsound

Key Functionalities:

Detects predefined emergency hand gestures Sends SOS alerts to pre-saved emergency contacts with live location Activates a sound alarm to alert nearby people Supports camera inputs from mobile or wearable devices

Features

Real-time gesture detection Continuous live location tracking Loud alarm activation Emergency contact list integration Works across web/mobile platforms

Challenges I Faced

Ensuring high gesture accuracy across different environments Preventing false alarms during normal hand movements Optimizing performance for real-time use Designing a user-friendly interface for quick activation under stress

Built With

• api
• css
• firebase
• geopy
• google
• html
• javascripttensorflow
• maps
• mediapipe
• mobile
• numpy
• opencv
• or
• pc
• playsoundruns-on-web-browser
• python
• smtplib