Aim

To train a simple computer model to recognize different hand gestures or objects using Google’s Teachable Machine.

Materials Required

• Computer or laptop with internet access

• Webcam

• Website: https://teachablemachine.withgoogle.com

• Chart paper for reflection (optional)

Step-by-Step Activity Plan

Step 1: Open Teachable Machine (5 min)

1. Visit teachablemachine.withgoogle.com.

2. Click “Get Started” → choose “Image Project.”

Step 2: Create Classes (5 min)

1. Create two classes — e.g.,

   • Class 1: “Hand Up”

   • Class 2: “Hand Down”
     (Students can choose different objects or gestures like “Book” vs “Pen” if available.)

2. Name your classes clearly.

Step 3: Capture Training Data (10 min)

1. Using the webcam, record 20–30 images per class.

2. Keep lighting and angles varied — sometimes bright, sometimes dim — to make the data diverse.

3. Click “Hold to Record” or “Capture” for each example.

Step 4: Train the Model (5 min)

1. Click “Train Model.”

2. Wait as the system processes your data.
   This step is where the computer “learns” to tell the difference between your two classes.

Step 5: Test the Model (10 min)

1. After training, use the preview window to test.

2. Move your hand up or down (or show your chosen object) and observe how AI responds.

3. Try changing the background or lighting — does it still recognize correctly?

4. If it makes mistakes, add more training images and re-train.