BINARY

Real-Time Violence Detection System

Overview

The Real-Time Violence Detection System is an AI-powered web application designed to detect violent activities in real time. The system uses deep learning models to analyze video feeds and alert authorities when violence is detected. Additionally, it features live location tracking for enhanced situational awareness.

Features

• Real-time violence detection using a trained deep learning model with our own built model using Conv3D, Maxpooling3D, Dense Layer etc.
• Web-based user interface for monitoring video feeds.
• Live location tracking for emergency response with alert sms to the contacts.
• High accuracy in detecting violent actions.

Tech Stack

• Frontend: HTML, CSS, JavaScript
• Backend: Python (Flask)
• Machine Learning Model: TensorFlow/Keras
• Additional Tools: OpenCV, W&B, YOLOv5

Model Download

Download the trained deep learning model (model.h5) from the following link: Click here to download

Installation

1. Clone the repository:

   https://github.com/Bipasha1005/BINARY.git
   cd violence-detection

2. Install dependencies:

   pip install -r requirements.txt

3. Place the model.h5 file in the models directory.
4. Run the application:

   python app.py

5. Open your browser and navigate to http://127.0.0.1:5000.

Usage

• Upload or stream a video feed.
• The system will analyze the frames in real time.
• If violence is detected, an alert will be triggered.
• Live location data will be sent to authorities (if enabled from the system).

Contributors

• Animesh Ghosh
• Anurag Kumar Thakur
• Bipasha Acharjee

Acknowledgments

• TensorFlow/Keras for deep learning
• OpenCV for image processing
• WebRTC for real-time video streaming