Model answer

Find the average color of each known class, then assign each pixel to whichever class average it’s closest to. That’s the whole rule.

Pros: - Fast. - Never leaves a pixel unclassified.

Cons: - Ignores class shape (a tight class and a loose class are treated identically). - Force-fits weird pixels that should probably be flagged as unclassified.

Good for a quick first pass. Use Max Likelihood if you want accuracy.

🔬 Show the science / technical version

• For each pixel, compute Euclidean distance to every class mean and assign to the closest.
• Strength: fast, works with small training sets.
• Weakness: ignores class shape/variance → poor on classes with different spreads or correlations.
• Everything gets classified (no unclassified pixels) — can force-fit outliers.