Sigima - Scientific Image and Signal Processing Library

Sigima is an open-source Python library for scientific image and signal processing, designed as a modular and testable foundation for building advanced analysis pipelines.

🔬 Developed by the DataLab Platform Developers, Sigima powers the computation backend of DataLab.

🚀 Try it Online

Experience Sigima instantly in your browser — no installation required!

Click the badge above to open a basic example notebook in a live JupyterLite environment powered by notebook.link. This service, developed by QuantStack, enables sharing and running Jupyter notebooks directly in the browser with zero setup.

Simply run the cells to explore:

• Creating signal and image objects
• Applying processing functions
• Visualizing results inline

🌟 Project & Sponsors

Project/Sponsor	Description
	Open-source platform for scientific signal and image processing, powered by Sigima.
	European non-profit supporting open-source and internet projects. Sigima has received funding from NLnet for its development, through the DataLab project.

✨ Highlights

• Unified processing model for 1D signals and 2D images
• Works with object-oriented wrappers (SignalObj, ImageObj) extending NumPy arrays
• Includes common processing tasks: filtering, smoothing, binning, thresholding, labeling, etc.
• Structured for testability, modularity, and headless usage
• 100% independent of GUI frameworks (no Qt/PlotPyStack dependencies)

💡 Use cases

Sigima is meant to be:

• A processing backend for scientific/industrial tools
• A library to build reproducible analysis pipelines
• A component for headless automation or remote execution
• A testbed for developing and validating new signal/image operations

📖 Design Philosophy

The main goal of Sigima is to provide a unified, high-level API for handling and processing 1D signals and 2D images, through dedicated Python objects: SignalObj and ImageObj.

The library is organized to separate concerns clearly:

• sigima.objects: defines the object model for signals and images.
• sigima.params: contains parameter classes for configuring processing functions.
• sigima.proc: provides high-level processing functions that operate directly on SignalObj and ImageObj instances.
• sigima.io: handles input/output operations (CSV files, image formats, etc.) for signals and images.
• sigima.tools: contains low-level, NumPy-based functions that implement the core logic behind many processing routines.

This structure supports a layered programming model:

• Developers can use computation to process full signal/image objects in an object-oriented manner.
• Or they can directly use tools to process raw NumPy arrays — for instance, in custom tools or when integrating Sigima into other projects.

⚠️ sigima.tools is not intended as a general-purpose NumPy extension. Its purpose is to fill in the gaps of common scientific libraries (NumPy, SciPy, scikit-image, etc.), offering consistent tools for signal/image processing in the context of Sigima and similar projects.

Usage Outside Sigima

Although Sigima is designed primarily for object-based processing, some of its core functions are useful on their own.

For instance, the DataLab project — an open-source platform for signal/image processing — uses many functions from sigima.tools independently of the object model. This demonstrates how sigima.tools can serve as a lightweight utility layer in scientific and industrial Python applications, even when the object model is not used directly.

To maintain this flexibility and avoid confusion, the distinction between tools (array-based) and computation (object-based) is intentional and explicit.

📦 Installation

Or in a development environment:

git clone https://github.com/DataLab-Platform/Sigima.git
cd Sigima
pip install -e .

📚 Documentation

📖 Full documentation (in progress) is available at: 👉 https://sigima.readthedocs.io/

Want to use Sigima inside DataLab with GUI tools? Check out the full platform: DataLab

⚙️ Architecture

Sigima is organized by data type:

sigima/
├── tools/      # Low-level NumPy-based algorithms supporting some computation functions
├── proc/       # High-level processing functions operating on SignalObj/ImageObj
│   ├── base/   # Common processing functions
│   ├── signal/ # 1D signal processing
│   └── image/  # 2D image processing

Each domain provides:

• Low-level functions operating on NumPy arrays
• High-level functions operating on SignalObj or ImageObj

🧪 Testing

Sigima comes with unit tests based on pytest.

To run all tests:

To run GUI-assisted validation tests (optional):

🧠 License

Sigima is distributed under the terms of the BSD 3-Clause license. See LICENSE for details.

🤝 Contributing

Bug reports, feature requests and pull requests are welcome! See the CONTRIBUTING guide to get started.

© DataLab Platform Developers