P5 Sketch Link: https://editor.p5js.org/my3037/full/_yVp2SasM

This project is an interactive real-time trainer that combines p5.js (for creative visuals and interaction) with ml5.js (for machine learning in the browser).

It lets you:

• Use Handpose to capture all 21 hand landmarks in real time.
• Collect data for different labels (classification) or RGB color values (regression).
• Train a Neural Network directly in the browser.
• See the model’s predictions instantly visualized as colors, shapes, and text.

You can even use mic loudness and device orientation as extra inputs — making the system responsive to gestures, sound, and motion at once.

Core Flow

1. Capture Input Data → hand keypoints, mic level, and orientation.

2. Add Labels or RGB Targets → decide what the model should learn.

3. Train the Neural Network → ml5 normalizes and learns from samples.

4. Predict in Real Time → the trained model outputs a label or RGB values.

5. Visualize the Result → the output drives background color or shapes.