SABERβοΈ β Segment Anything Based Electron Tomography Recognition
Welcome to the documentation for SABER, a robust, open-source platform for autonomous segmentation of organelles in cryo-electron tomography (cryo-ET) and electron microscopy (EM) datasets.
SABER leverages state-of-the-art foundational models to enable:
- Zero-shot segmentation of EM/cryo-ET data
- Interactive annotation with a user-friendly GUI
- Expert-driven classifier training for improved accuracy
- 3D organelle reconstruction from tomographic data
π What Can You Do With SABER?
- Pre-process and curate your microscopy data for segmentation
- Train custom classifiers using your own expert annotations
- Run inference for both 2D and 3D segmentation tasks
- Visualize and refine results interactively
π‘ Why SABER?
SABER is designed for researchers who want to:
- Accelerate organelle segmentation in large EM/cryo-ET datasets
- Combine the power of foundational models with expert curation
- Seamlessly move from raw data to publication-ready segmentations in both 2D and 3D.
π Tutorials
SABER can be used both from the command line (CLI) and as a Python library (API). Choose the workflow that fits your needs. Follow the quick start guide to learn the core commands to develop segmentations.
π₯οΈ CLI Tutorials
Use SABER from the command line for quick, scriptable workflows:
- Pre-processing Your Data: Prepare your EM/cryo-ET datasets for segmentation and annotation. Use the interactive GUI to annotate segmentations and
- Training a Classifier: With the annotations, train a domain expert classifier.
- Inference in 2D & 3D: Apply your trained models to generate high-quality segmentations in both 2D and 3D.
- Refine Membranes: Clean up organelle and membrane segmentations using GPU-optimized morphological filtering.
π API Tutorials
Integrate SABER into your own Python scripts and notebooks:
- API Overview: Comprehensive introduction to the SABER Python API
- 2D Quickstart: Learn how to segment 2D micrographs programmatically
- 2D Workflow Tutorial: Complete 2D pipeline with preprocessing, training, and inference
- 3D Quickstart: Learn how to segment 3D tomograms programmatically
- 3D Workflow Tutorial: Complete 3D pipeline with video propagation and advanced features
πββοΈ Getting Help
Visit our GitHub repository for source code and issues.
Ready to get started? Check out the Quick Start or the API Overview!