RAGE Detector

The current models are trained using my own data. Guide to train your own is in the Training Your Own Model Section

A real-time emotion detection system designed to identify rage and anger in gaming contexts, particularly for reaction to bad randoms in Marvel Rivals and other competitive games.

This project uses computer vision and machine learning to analyze facial expressions, body posture, and hand gestures in real-time to detect when a player is experiencing rage or anger, or happiness when hitting a 3-man Ironman Ult. The system leverages MediaPipe for pose and facial landmark detection, combined with trained machine learning models to classify emotional states.

Installation

1. Clone the repository

   git clone <repository-url>
   cd Rage-Detector

2. Create a virtual environment (recommended)

   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install dependencies

   pip install -r requirements.txt

Running the Rage Detector

1. Start the main application

   cd main
   python main.py

2. Use the application

   • The webcam feed will open with real-time landmark detection
   • Emotion classification results will be displayed on screen
   • Press 'q' to quit the application

Training Your Own Model

1. Collect training data

   cd training
   python collectData.py

   • Modify the current_feeling variable in collectData.py to collect different emotions
   • Set csv_initialized = False to start fresh data collection
   • The system will capture pose and facial landmarks while you express different emotions
   • ***Remember to set csv_initialized = True after initializing the file once

2. Train the model

   python train.py

   • This will train multiple machine learning models on your collected data
   • Models are saved as pickle files for later use

3. Handle CSV data (optional)

   python handleCSV.py

   • Use this script to process and clean your training data

Project Structure

Rage-Detector/
├── main/                   # Main application files
│   ├── main.py            # Real-time detection application
│   └── getResult.py       # Result processing utilities
├── training/              # Model training and data collection
│   ├── collectData.py     # Data collection script
│   ├── train.py          # Model training script
│   ├── handleCSV.py      # CSV data processing
│   ├── data.csv          # Training dataset
│   ├── LR.pkl            # Logistic Regression model
│   ├── RF.pkl            # Random Forest model
│   └── GB.pkl            # Gradient Boosting model
├── requirements.txt       # Python dependencies
├── LICENSE               # Project license
└── README.md            # This file

Model Selection

The system currently uses Logistic Regression by default. Use other models by:

1. Uncomment the desired model in training/train.py
2. Retrain the model
3. Update the model loading in main/main.py

HAPPY RAGING BECAUSE OF BAD RANDOMS!!!