Image Vectorization with Python

Convert raster images into stylized, vector-like artwork using machine learning techniques.

🖼️ Before & After Examples

Original	Vectorized

The example room images used in this repository were created by me using 3D modeling in Blender for this project.

🎓 Project Background

Developed for Application Design 2 course (4th semester Interaction Design) at HfG Schwäbisch Gmünd. Created to vectorize image sequences with consistent colors for animations.

🔗 University Project: SideBy - HfG Exhibition

✨ Key Features

• 🎨 Color Reduction via K-means clustering
• ⚫ Black & White Mode with customizable grey levels
• 🎬 Sequence Processing for consistent animation colors
• 📝 Edge Detection for manga/anime style outlines
• ⚙️ Easy Configuration - just edit settings at top of files

🎨 Style Gallery

Manga Style	Minimalist	Black & White
Thin Lines	No Edges	Simple B&W

�🚀 Quick Setup

pip install opencv-python numpy matplotlib scikit-learn scipy pillow
python examples.py  # See style gallery

🎯 Usage

1. Explore Styles

python examples.py

Check examples/style_gallery/ for different artistic styles.

2. Single Images

Edit settings in image_vectorizer.py:

DEFAULT_COLORS = 8         # Number of colors (2-50)
DEFAULT_BLACK_WHITE = False # True for B&W mode
DEFAULT_EDGE_THRESHOLD1 = 100  # Edge thickness

Then run: python image_vectorizer.py

3. Image Sequences (For animations)

Put images in examples_image_sequence/room-animation/ Edit sequence_vectorizer.py:

SEQUENCE_COLORS = 6        # Colors for entire sequence
SEQUENCE_BLACK_WHITE = False # True for B&W sequences

Then run: python sequence_vectorizer.py

4. Command Line

python main.py --input "photo.jpg" --output "result.png" --colors 6 --black-white

🎨 Style Examples

See the Style Gallery above for visual examples of each style:

Manga Style: 6-8 colors, edges 30/100, B&W mode optional
Minimalist: 3-4 colors, smoothing 3.0
Clean Lines: 8-12 colors, edges 80/200

� Main Files

• image_vectorizer.py - Main engine, edit settings here
• sequence_vectorizer.py - For consistent animation sequences
• examples.py - Style gallery generator
• main.py - Command line interface

� Key Settings

• Colors: 3-6 (minimal), 8-12 (balanced), 15+ (detailed)
• Black & White: Use 4-10 colors for grey levels
• Edges: 30/100 (thick manga), 80/200 (thin), 100/250 (minimal)

License

MIT License - Feel free to use and modify for your own projects!