Week 1 · Day 1

How a Computer Sees

Spoiler: a photo is just a giant grid of numbers. Today you'll zoom inside a picture, mix light like a screen, split colors apart, and do real image math — the exact stuff every AI uses to see.

↓ scroll to start playing
🏷️ The pro wordA digital image is a grid of pixels, each holding Red, Green, Blue values (0-255). Grown-ups call that grid of numbers a tensor, and the R/G/B layers are channels. Knowing these words lets you talk to any image library.
🔍 Toy 1 · Zoom In

A photo is made of tiny squares

Move your mouse (or finger) over the photo. The magnifier shows the real pixels underneath — and the three numbers hiding inside each one.

Hover the photo…
Pixel = one tiny square of the image. Each one stores 3 numbers: how much Red, Green and Blue light it glows — from 0 (off) to 255 (full blast).

Peek at 15 different spots to earn the badge. Peeked: 0

🔍 Curious? What's a "resolution"?

Resolution just means how many pixels the grid has. More squares = sharper photo = a bigger pile of numbers. A phone photo can be 12 million pixels — that's 36 million numbers! Grown-ups call this grid of numbers a tensor.

🎨 Toy 2 · Mix Light

Every color = 3 numbers

Computers make color by mixing light, not paint. Drag the sliders and watch three beams of light blend. This is called RGB.

220
60
90
rgb(220, 60, 90)
🎯 Challenge: Make bright yellow ☀️

💡 Surprise: Red + Green light = Yellow. All three at max = White. All zero = Black.

🧩 Toy 3 · Split It

One photo is really 3 photos stacked

Every image is three layers of numbers — a Red layer, a Green layer, and a Blue layer — stacked on top of each other. Tap a layer to switch it off and watch the color change.

🌈 Full color
Red layer
Green layer
Blue layer
Bright = a big number, dark = a small number. Where the Red layer is bright, that spot has lots of red in it. Turn layers off and see how the full-color photo changes!

Toggle all 3 layers at least once to earn the badge.

🧮 Toy 4 · Do the Math

Editing a photo = simple math on the numbers

Watch what happens to one example pixel when you press a button. There's no magic — filters are just tiny sums, done to every single pixel at once.

🔬 The math on one pixel:

Pick an operation to see the math…
Big idea: A computer does this same little sum to every pixel — thousands of times in a blink. That's all an image "filter" really is.
🔍 Curious? What does grayscale mix mean?

We don't just average R, G, B equally. Our eyes see green as brightest and blue as dimmest, so the recipe is 0.3×R + 0.59×G + 0.11×B. That's why a good grayscale looks natural and not muddy.

🧙 Boss Build · Make Something

Speak the computer's language

You've learned that pictures are just numbers. Now you be the artist: click cells to paint an 8×8 pixel emoji. Every square you fill is you setting numbers by hand — exactly how the computer stores it.

Try this: make a heart ❤️, a smiley 🙂, or a ghost 👻. Fill at least 8 squares, then hit Finish to earn your Pixel Wizard badge.

Every colored square = 3 numbers you chose. You just hand-wrote an image the way a computer stores one. 🤯

🚀 Beyond the Basics

How the pros really do it

Reading raw RGB is the foundation. Real vision systems add a few ideas on top:

🟢

Raw RGB (what you learned)

✅ Simplest and exact. ⚠️ Big files; not always the best form for a task.

🎨

Color spaces

Engineers switch to grayscale, HSV, or YUV depending on the job. ✅ Can make patterns much easier to spot.

🧠

How AI uses pixels

Models normalize the pixels, then a CNN learns its own features from them — you rarely pick pixels by hand.

🏷️ Pro names to look uptensor · channels · color spaces (RGB / HSV / grayscale) · normalization · formats (JPEG / PNG).

🎉 Day 1 Complete!

You now know the secret every AI is built on: a picture is a grid of numbers, and seeing is just math.

Day 2: Find the Edges →
Tomorrow: we take these number-grids and slide a little "magic window" over them to detect edges — the exact trick that lets AI find shapes, faces, and objects. 👀
AI Adventure · Week 1 · Day 1 — made for young inventors 🚀