Image Compression is a modern Flask web app that reduces image size using K-Means color quantization.

Instead of storing millions of unique colors, the algorithm intelligently groups similar colors into clusters — dramatically shrinking file size while preserving visual quality.

📦 From a simple Jupyter Notebook experiment, this project has been upgraded into a fully interactive web application with:

• ✔ Upload
• ✔ Compress
• ✔ Compare
• ✔ Download
• ✔ View metrics

All in seconds.

Upload Image
      ↓
K-Means Clustering (Color Quantization)
      ↓
Rebuild using K colors
      ↓
Show Preview + Compression Stats
      ↓
Download Optimized Image

• K-Means color compression
• Adjustable cluster count (8 – 64 colors)
• RGB → cluster → reconstruct pipeline
• Automatic JPEG optimization

• Unique original colors count
• Original file size
• Compressed file size
• Compression percentage
• Side-by-side comparison

• Glassmorphism design
• Gradient theme
• Smooth animations
• Loading state
• Mobile responsive
• Bootstrap powered

• Flask routing
• Static file handling
• Secure uploads
• Fast NumPy processing

Image → reshape to:

(height × width, 3)

Each pixel becomes:

[R, G, B]

Group pixels into K clusters

Example:

16 million colors → reduced palette ( 8, 16, 32, 64 )

Replace each pixel with its cluster center.

Result:

Smaller memory + Similar visual quality

image-compression/
│
├── app.py
├── requirements.txt
│
├── static/
│   ├── uploads/
│   └── outputs/
│
├── templates/
│   └── index.html
│
└── README.md

git clone https://github.com/SACHIN-S-2004/Image-Compression.git
cd Image-Compression

pip install -r requirements.txt

python app.py

http://127.0.0.1:5000

• Drag & drop upload
• Multiple images batch compression
• Slider for real-time K selection
• React frontend version
• Auto cluster suggestion

This project demonstrates:

• ✔ Unsupervised Learning (K-Means)
• ✔ Image processing fundamentals
• ✔ Flask backend development
• ✔ Practical ML deployment

Give it a star — it helps a lot!