Video and Audio Feature Extraction Project

This project processes video and audio files to extract various features such as pose landmarks, facial blink ratio, audio tempo, and pitch. It also simulates skin signal changes for further analysis.

Features Extracted

1. Pose Features:

   • Head tilt based on the distance between left and right eyes.
   • Body movement based on the distance between left and right shoulders.

2. Facial Features:

   • Blink ratio calculated using MediaPipe's FaceMesh.

3. Audio Features:

   • Tempo and average pitch extracted using librosa.

4. Simulated Skin Signal:

   • Generated sinusoidal waves to simulate skin signal changes.

Files

• video_path: Input video file to be processed.
• audio_output_path: Path for the extracted audio file from the video.
• body_language_features.csv: Output CSV containing the processed features.

Requirements

The following Python libraries are required:

• cv2 (OpenCV)
• mediapipe
• numpy
• pandas
• moviepy
• librosa
• matplotlib

Install these using:

pip install opencv-python mediapipe numpy pandas moviepy librosa matplotlib


---

## Contributors & Domain Experts

<table>
  <tr>
    <td align="center">
      <a href="https://github.com/mlnjsh">
        <img src="https://github.com/mlnjsh.png" width="80px;" alt="Milan Amrut Joshi"/><br />
        <sub><b>Milan Amrut Joshi</b></sub>
      </a><br />
      <sub>Project Author</sub>
    </td>
    <td align="center">
      <a href="https://github.com/ultralytics">
        <img src="https://github.com/ultralytics.png" width="80px;" alt="Ultralytics"/><br />
        <sub><b>Ultralytics</b></sub>
      </a><br />
      <sub>YOLOv8 object detection</sub>
    </td>
    <td align="center">
      <a href="https://github.com/open-mmlab">
        <img src="https://github.com/open-mmlab.png" width="80px;" alt="OpenMMLab"/><br />
        <sub><b>OpenMMLab</b></sub>
      </a><br />
      <sub>MMDetection, MMTracking</sub>
    </td>
  </tr>
</table>