NeuroFeel Documentation

Cross-Dataset Emotion Recognition Framework Documentation

Installation

Setting up the Cross-Dataset framework

git clone https://github.com/HimashaHerath/NeuroFeel.git
cd neurofeel
pip install -r requirements.txt

This will install all required dependencies including NumPy, Pandas, Scikit-learn, and PyTorch.

Environment Variables

Set the following environment variables to point to your datasets:

export WESAD_PATH=/path/to/wesad/dataset
export KEMOCON_PATH=/path/to/kemocon/dataset
export RESULTS_DIR=/path/to/output/directory

Framework Overview

Key components and structure

The Cross-Dataset Emotion Recognition Framework consists of two main components:

WESAD Framework

Personalized emotion recognition for WESAD dataset with adaptive model selection.

Cross-Dataset Framework

Bidirectional model transfer between WESAD and K-EmoCon with domain adaptation.

Directory Structure

neurofeel/
├── wesad_framework/      # WESAD personalization framework
├── cross_dataset/        # Cross-dataset framework
│   ├── data/             # Data loaders
│   ├── features/         # Feature extraction
│   ├── domain_adaptation/# Domain adaptation methods
│   ├── models/           # Model training
│   ├── evaluation/       # Evaluation utilities
│   ├── visualization/    # Visualization tools
│   ├── config.py         # Configuration
│   ├── framework.py      # Main framework class
│   └── main.py           # Command-line interface
├── api/                  # API implementations
│   ├── wesad_api.py      # WESAD framework API
│   └── cross_dataset_api.py # Cross-dataset API
└── requirements.txt      # Dependencies

Resources

GitHub Repository WESAD Framework Demo Cross-Dataset Framework Demo