🖼️ Day 8

The 3 Flavors of Vision

There are three big ways an AI can "see" a picture — from a quick label, to boxes, to coloring in every single pixel. Let's line them up side by side and feel the difference.

↓ let's compare
🏷️ The pro wordLabeling every pixel is segmentation: semantic (color each pixel by class), instance (separate each object), and panoptic (both).
👁️ The Three Ways to See

Same picture, three answers

🎨 Lab · Paint the Pixels

Segmentation = color every pixel

Detection draws a box (fast, but grabs background too). Segmentation is way pickier: it labels every single pixel as "object" or "not". Click and drag to paint the cat's pixels. Cover the cat, avoid the background — your overlap score climbs!

Overlap with the true cat mask (IoU)
0%
🎯 Challenge: get 80%+ overlap. This is exactly how apps blur your video-call background or make stickers — they segment YOU from everything else. 🎥
🧠 Boss Build · Pick the Right Tool

Which flavor fits the job?

Real engineers pick the right kind of vision for each task. Read each mission and choose!

Score: 0 / 0
🔬 Real Model Lab

Make up your own labels — real CLIP

Here's the mind-blowing one. CLIP is a real AI that matches pictures to words — and you pick the words! Type any labels you invent ("a happy dog", "my homework", "a toy robot") and it scores your photo against them. No retraining, ever. Running on your device. (~150 MB one-time — the biggest model here, so it's best on a laptop with a bit of room.)

not loaded
🚀 Beyond the Basics

How the pros really do it

You painted a mask by hand. Real segmentation uses trained networks:

🟢

Paint pixels by hand (what you did)

✅ Shows what a mask is. ⚠️ Way too slow for real use.

🖼️

Segmentation networks

U-Net, DeepLab, Mask R-CNN - trained to mask objects automatically.

🧠

Foundation models (AI)

SAM (Segment Anything) cuts out almost any object with a click. ✅ Magical. ⚠️ Big.

🏷️ Pro names to look upsemantic / instance / panoptic segmentation · U-Net / DeepLab / Mask R-CNN · SAM.

🎉 Day 8 Complete!

You now know the three flavors of computer vision — classification (what), detection (what + where, boxes), and segmentation (every pixel) — and when to use each.

Day 9: How YOLO Works →
Tomorrow: the famous YOLO detector — "You Only Look Once." You'll see how it spots every object in a picture in a single glance, super fast. ⚡
AI Adventure · Module 1 · Day 8 — made for young inventors 🚀