Gait Analysis to Detect Emotion

This project involves using gait analysis to detect human emotions. It leverages machine learning techniques to analyze human gait data and determine emotional states.

Table of Contents

• Introduction
• Features
• Installation
• Usage
• Directory Structure
• Contributing
• References

Introduction

The goal of this project is to analyze human gait data and detect emotions based on the analysis. This can have applications in security, healthcare, and entertainment.

Features

• Data extraction using OpenPose
• Data preprocessing and conversion
• Machine learning model training and evaluation
• Visualization of results

Installation

1. Clone the repository:

   git clone https://github.com/sarojshakya01/gait-emotion.git
   cd gait-emotion

2. Install the required dependencies:

   pip install -r requirements.txt

Usage

1. Extract data using OpenPose:

   python extract_openpose.py

2. Convert data formats:

   python h5_to_csv.py
   python h5_to_npy.py

3. Train the model:

   python main.py

4. Visualize the results:

   python animate_data.py

Directory Structure

• net/: Contains the neural network models.
• test/: Contains test scripts and datasets.
• utils/: Utility scripts for data processing and visualization.
• animate_data.py: Script for visualizing data.
• extract_openpose.py: Script for extracting data using OpenPose.
• h5_to_csv.py: Script for converting H5 data to CSV format.
• h5_to_npy.py: Script for converting H5 data to NPY format.
• main.py: Main script for training the model.

Contributing

Contributions are welcome! Please fork the repository and create a pull request with your changes.

Reference

Paper

https://arxiv.org/abs/1910.12906

Code

https://github.com/UttaranB127/GeneratingEmotiveGaits

---

This README provides an overview and guidance for using the repository effectively.