Welcome to Adorym’s documentation!

Adorym, (Automatic Differentiation-based Object Reconstruction with Dynamical Scattering) is a Python package for solving inverse problems in generic optical image reconstruction tasks (e.g., coherent diffraction imaging, holography, ptychography, line-projection tomography, ptychotomography, and multislice ptychotomography). The package uses automatic differentiation to calculate the gradient used for iterative optimization algorithms from the given forward model with minimal human intervention, saving the labors needed for rederiving the gradient when one switches/tweaks the forward model. Adorym supports the use of user-defined forward model, making it easy to test/build algorithms for novel imaging techniques.

Contents:

• Installation
• Quick start guide
  • Running a demo script
  • Running your own jobs
• Usage
  • Dataset format
  • Parameter settings in main function
  • Forward models
  • Regularizers
  • Optimizers
  • Output
• Customization
  • Adding your own forward model
  • Adding refinable parameters
• API references
  • adorym.array_ops module
  • adorym.constants module
  • adorym.conventional module
  • adorym.differentiator module
  • adorym.forward_model module
  • adorym.regularizers module
  • adorym.global_settings module
  • adorym.linesearch module
  • adorym.misc module
  • adorym.optimizers module
  • adorym.propagate module
  • adorym.pseudo module
  • adorym.ptychography module
  • adorym.simulation module
  • adorym.util module
  • adorym.visualization module
  • adorym.wrappers module
  • Module contents
• Publications