Tomocam-lamino Documentation

Tomocam-lamino is a high-performance C++ library for tomographic reconstruction of magnetic fields in thin materials exhibiting magnetic circular dichroism (MCD). Developed at Lawrence Berkeley National Laboratory, it provides advanced reconstruction algorithms optimized for parallel computing.

Contents:

• Getting Started
  • What is Tomocam-lamino?
  • Prerequisites
  • Key Concepts
  • First Steps
  • Next Steps
• Installation
  • Requirements
  • Installing Dependencies
  • Building Tomocam-lamino
  • Building with Tests
  • Installation
  • GPU Support (Optional)
  • Verifying Installation
  • Troubleshooting
• Usage Guide
  • Command Line Interface
  • Workflow Overview
  • Input Data Format
  • Reconstruction Types
  • Basic Example
  • Advanced Usage
  • Output Formats
  • Visualization
  • Performance Tips
  • Troubleshooting
  • Next Steps
• Configuration Reference
  • Configuration File Structure
  • Input Section
  • Output Section
  • Reconstruction Parameters
  • Regularization
  • Complete Example
  • Parameter Selection Guide
  • Best Practices
  • Troubleshooting
• API Documentation
  • Core API
  • Reconstruction API
  • I/O API
  • Utilities API
  • Core Components
  • Reconstruction Algorithms
  • I/O Operations
  • Utilities
  • Class Reference
  • Function Reference
  • Type Definitions
  • Usage Example
  • Build Configuration
  • See Also
• Examples
  • Example 1: Basic Scalar Reconstruction
  • Example: Vector Field Reconstruction
  • Example 3: High-Resolution with qGGMRF
  • Example 4: Batch Processing
  • Example 5: Forward Projection (Simulation)
  • Example 6: Python Integration
  • Example 7: ParaView Visualization
  • Next Steps
• Contributing
  • Getting Started
  • Development Setup
  • Code Style
  • Documentation
  • Testing
  • Pull Request Process
  • Code Review
  • Reporting Issues
  • Community
  • License
  • Questions?
• License
  • Copyright Notice
  • U.S. Government Rights
  • Contact Information
  • Full License Text

Overview

Tomocam-lamino provides:

• Forward/Backward Projection: Efficient projection operators on polar grids

• Iterative Reconstruction: Conjugate gradient and Nesterov accelerated gradient methods

• MBIR: Model-Based Iterative Reconstruction with QGGMRF penalty

• TIFF I/O: Read and write reconstruction data

• GPU Acceleration: Optional CUDA support for enhanced performance

Features

Performance Optimizations

• OpenMP parallelization for multi-core CPUs

• Intel TBB (Threading Building Blocks) for task-based parallelism

• FFTW for fast Fourier transforms

• FINUFFT for non-uniform FFTs

• Optional GPU acceleration via CUDA

Supported Platforms

• Linux (tested on Arch Linux)

• macOS (15.5+)

Quick Start

Build the project:

# Linux
cmake --preset arch
cmake --build --preset arch

# macOS
cmake --preset macos
cmake --build --preset macos

Run reconstruction:

./build/recon config.toml

Indices and tables

• Index

• Search Page