python 33 lines · 7 steps

Extracting a color palette with K-means

Cluster an image's pixels into a handful of dominant colors and rank them by how much of the image they cover.

Explained by highlit
Intermediate clustering image-processing kmeans numpy color
1from collections import Counter
2 
3import numpy as np
4from PIL import Image
5from sklearn.cluster import KMeans
6 
7 
8def extract_palette(image_path, num_colors=6, resize_to=200, sample_frac=1.0):
9 image = Image.open(image_path).convert("RGB")
10 image.thumbnail((resize_to, resize_to))
11 
12 pixels = np.asarray(image, dtype=np.float64).reshape(-1, 3)
13 if sample_frac < 1.0:
14 rng = np.random.default_rng(seed=42)
15 idx = rng.choice(len(pixels), size=int(len(pixels) * sample_frac), replace=False)
16 pixels = pixels[idx]
17 
18 kmeans = KMeans(n_clusters=num_colors, n_init="auto", random_state=42)
19 labels = kmeans.fit_predict(pixels)
20 
21 counts = Counter(labels)
22 total = sum(counts.values())
23 centers = kmeans.cluster_centers_.round().astype(int)
24 
25 palette = []
26 for label, count in counts.most_common():
27 r, g, b = centers[label]
28 palette.append({
29 "hex": f"#{r:02x}{g:02x}{b:02x}",
30 "rgb": (int(r), int(g), int(b)),
31 "weight": round(count / total, 4),
32 })
33 return palette
01 / 01
STEP 01

Walkthrough

Space play ←→ step click any line
Three takeaways
  1. 1Clustering pixel colors turns thousands of shades into a few representative swatches.
  2. 2Downsizing and optionally subsampling keeps the clustering fast without changing the palette much.
  3. 3Cluster sizes double as weights, letting you rank colors by their visual prominence.

Related explainers

python 
from fastapi import FastAPI, WebSocket, WebSocketDisconnect
 
app = FastAPI()
 

Building a WebSocket chat with FastAPI

websockets broadcast connection-management
Intermediate 9 steps
python 
import time
import uuid
 
from django.utils.deprecation import MiddlewareMixin

Attaching per-request context in Django

middleware request lifecycle multi-tenancy
Intermediate 7 steps
python 
import random
from typing import Iterator, List
 
 

How reservoir sampling picks k items

reservoir-sampling streaming randomness
Intermediate 5 steps
python 
import secrets
 
from django.contrib.auth import authenticate, login
from django.core.cache import cache

Two-factor login with OTP in Django

two-factor-auth one-time-passwords caching
Intermediate 9 steps
python 
import re
from functools import total_ordering
from typing import Optional
 

Parsing and comparing semantic versions

regex operator-overloading sorting
Intermediate 7 steps
python 
from typing import Any, Sequence, Mapping
 
 
def render_markdown_table(

Rendering an aligned Markdown table in Python

string formatting data transformation closures
Intermediate 8 steps

Share this explainer

Here's the card — post it anywhere.

Extracting a color palette with K-means — share card
Made with highlit — turn any snippet into a walkthrough like this in about a minute.
Explain your code