Skip to content

Engagement Analysis with Head Pose Estimation is a computer vision project that utilizes Mediapipe library for facial landmarks detection, OpenCV for computer vision tasks, and NumPy/Pandas for data manipulation. It estimates head pose and gaze direction to determine whether the user is engaged or not.

License

Notifications You must be signed in to change notification settings

Manoj-2702/EngageVision

Repository files navigation

EngageVision: AI-Powered Attention

logo

Engagement Analysis with Head Pose Estimation

Engagement Analysis with Head Pose Estimation is a computer vision project that utilizes Mediapipe library for facial landmarks detection, OpenCV for computer vision tasks, and NumPy/Pandas for data manipulation. It estimates head pose and gaze direction to determine whether the user is engaged or not.

Features

  • Head pose estimation.
  • Gaze direction analysis.
  • Engagement analysis based on head pose and gaze direction.
  • Real-time visualization of facial landmarks and analysis results.
  • Track the duration of continuous eye contact with the camera to measure attention span.
  • Analyze variations in gaze direction to understand shifts in focus.

Parameters

Adjustable parameters are available in the script for customization:

  • draw_gaze: Set to True to display gaze vectors.
  • draw_full_axis: Set to True to display the full rotation axis.
  • draw_headpose: Set to True to display head pose vectors.
  • x_score_multiplier and y_score_multiplier: Multipliers to adjust the impact of gaze on head pose estimation.
  • threshold: Threshold for averaging gaze scores between frames.

Installation

  1. Clone the repository:

    git clone https://github.com/Manoj-2702/EyeTracking.git
    
  2. Install all the required Dependencies: Requirements:

    • OpenCV
    • Mediapipe
    • Numpy
    • Pandas
    pip install -r requirements.txt
    
  3. Usage:

    python eye_nose_detector.py
    

Results

The script provides real-time visual feedback on engagement and gaze direction. It also generates a summary of engagement statements and percentages after execution.

Sample Result

image

Engagement.Analysis.2023-11-12.20-22-11.mp4

License

This project is licensed under the MIT License - see the LICENSE file for details.

About

Engagement Analysis with Head Pose Estimation is a computer vision project that utilizes Mediapipe library for facial landmarks detection, OpenCV for computer vision tasks, and NumPy/Pandas for data manipulation. It estimates head pose and gaze direction to determine whether the user is engaged or not.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages