🧑‍💻 Online Body Data Collection System

This project is an online system for collecting body data using Vue3 as the frontend framework and Express as the backend. It leverages Mediapipe models for real-time body posture recognition and uses Axios to send the body coordinates to the server.

🖥️ Frontend: Vue

The frontend is developed using Vue3 and includes real-time body posture recognition. It utilizes:

• Mediapipe Models:
  • BlazePose
  • HandPose

These models provide normalized body coordinates (ranging from 0 to 1), which are sent to the server for further processing. Invalid data is identified by coordinates less than 0 or greater than 1.

📡 Backend: Express

1. MongoDB Connection: Found in express/app.js
2. API Requests: URL endpoints are configured in vue/src/api/axios.js

Body Data

The collected body data is normalized, with coordinates ranging from 0 to 1. To convert these into actual pixel values for a video frame, use the following formula:

widthPx = x * width
heightPx = y * height

• Width: 1280 pixels
• Height: 720 pixels

File Interpretation

• /express/uploads/: Contains audio files named in the format username_dataid_recordingnumber.mp3, e.g., 1840_10_1.mp3 means the audio file is the first recording by user 1840 during data recording session ID 10.

• /express/bodyData/: Contains JSON files named in the format username_dataid.json, e.g., 2107_6.json represents the body data collected by user 2107 during data recording session ID 6.

🎥 Video Recall Script

A script is available to reconstruct the collected body data from JSON format.

---

Feel free to customize the README further as needed. If you have any additional details or want specific sections expanded, let me know!