Build a Teachable Machine Image Classifier With Your Kid

Introduce your child to the fundamentals of machine learning through a hands-on, interactive web project.

Demystifying artificial intelligence for the next generation does not require complex coding bootcamps or expensive software. Using a free web tool and a standard laptop webcam, you and your child can build a functional machine learning model in fifteen minutes.

Children between the ages of 8 and 16 are surrounded by AI, from video game recommendation engines to facial recognition on tablets. However, they rarely get to see behind the curtain. By building a custom image classifier, they transition from passive consumers of technology to active creators, gaining a foundational understanding of how computers learn to see.

The Project: Rock, Paper, Scissors Classifier

We will build a model that can instantly recognize whether a hand is showing "Rock," "Paper," or "Scissors." This classic game provides a perfect framework because the hand shapes are distinct, easy to capture, and highly visual.

No software installation or coding experience is required. The entire project runs directly in your web browser on any standard laptop or desktop computer equipped with a webcam.

Step 1: Accessing the Platform

To begin, open your web browser and navigate to the official Google Teachable Machine website:

https://teachablemachine.withgoogle.com/

Click on the Get Started button, then select Image Project from the available options. Choose the Standard Image Model . This opens the main workspace, which is divided into three clear sections: Gathering Data (Classes), Training, and Preview.

Step 2: Gathering and Labeling Training Data

In machine learning, a "Class" is a category of things we want the computer to recognize. We need to create three distinct classes for our game:

• Class 1 (Rock): Rename "Class 1" to "Rock". Click the Webcam button. Have your child hold up a fist in front of the camera. Click and hold the Hold to Record button to capture about 100 to 150 images. Encourage them to rotate their hand, move it closer and further away, and tilt it slightly. This variety helps the model learn the shape of a fist rather than just one specific angle.
• Class 2 (Paper): Rename "Class 2" to "Paper". Click the Webcam button. Have your child hold up an open, flat hand. Record another 100 to 150 images, varying the hand position and distance.
• Class 3 (Scissors): Click Add a class at the bottom. Rename it to "Scissors". Have your child hold up two fingers in a "V" shape. Record 100 to 150 images with similar variations.

Step 3: Training the Model

Once you have gathered data for all three classes, look at the middle column labeled Training . Click the blue Train Model button.

During this step, the browser processes the images and identifies patterns that distinguish a fist from an open hand or a "V" shape. Do not close the browser tab or switch to another window while the training is in progress, as this process occurs entirely within your local browser memory.

Step 4: Testing and Iterating

Once training is complete, the Preview column on the right will activate, showing a live feed from your webcam. Have your child make a fist, an open hand, or a "V" shape in front of the camera.

Watch the output bars below the video feed. They will dynamically shift from 0% to 100% based on what the model thinks it is seeing. If the model struggles to distinguish between "Paper" and "Scissors," discuss why this might be happening. Is the background too busy? Was the lighting different? You can easily add more images to any class and retrain the model to improve its accuracy.

Parent & Educator Sidebar

Core AI Concept: Supervised Learning This project demonstrates supervised learning, where a computer is trained using labeled examples (the images we captured). The computer does not "know" what a hand is; instead, it analyzes pixel patterns to find mathematical similarities between the training images and the live webcam feed.

Conversation Starters:

• "What do you think would happen if we trained the model with my hand, and then you tried to play? Why might it get confused?"
• "How could we teach this model to recognize a fourth option, like a thumbs-up gesture?"
• "Why is it important to have different backgrounds and lighting when we take the training photos?"

Action Steps Summary

• Open the Tool: Go to https://teachablemachine.withgoogle.com/ on a laptop.
• Capture the Data: Record 100 to 150 webcam images for Rock, Paper, and Scissors.
• Train the Model: Click "Train Model" and keep the browser tab open.
• Test and Refine: Use the live preview to test the model and discuss the underlying concepts.

Looking for more engaging ways to introduce your family to the world of technology?