Milenilines

Milenilines converts an image into a smooth vector field, traces dense streamlines, and automatically fits the result to a canonical ODE flow model.
The pipeline is Python-driven, with performance-critical components implemented in Rust.

---

Features

• Image → scalar potential field
• Gradient-based velocity field
• Dense streamline seeding and tracing
• Automatic ODE model selection via scoring
• Streamline morphing animation:
  • image only
  • image + streamlines
  • image streamlines → ODE streamlines
• Fast backend using Rust

---

Repository structure

.
├── app.py
├── requirements.txt
├── examples/                # licensed example images
├── milenilines/             # Python implementation
│   ├── io.py
│   ├── potential.py
│   ├── velocity.py
│   ├── streamlines.py
│   ├── render.py
│   └── ode_library.py
└── milenilines_rs/          # Rust (PyO3) acceleration module
    ├── Cargo.toml
    ├── Cargo.lock
    ├── pyproject.toml
    └── src/lib.rs

---

Installation

1. Create and activate virtual environment

python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

2. Build Rust extension (development mode)

cd milenilines_rs
export PYO3_USE_ABI3_FORWARD_COMPATIBILITY=1
maturin develop
cd ..

---

Usage

python app.py examples/name.jpg

---

Example images & licensing

Example images in examples/ are free to use and redistribute under.

• Unsplash (Unsplash License)

---

License

MIT License
See LICENSE for details.