Style Transfer Pipeline for Synthetic Cryo-ET Data

A comprehensive pipeline for applying style transfer to synthetic cryo-electron tomography (cryo-ET) data using faket, with support for micrograph projection, reconstruction, and training data preparation.

Overview

This pipeline processes cryo-ET simulation data and applies neural style transfer to generate augmented training datasets. It handles:

• Style micrograph projection from tomograms
• Clean and noisy micrograph generation
• Neural style transfer using faket
• Tomogram reconstruction from style-transferred micrographs
• Training data organization for downstream tasks

Requirements

Core Dependencies

• Python 3.7+
• NumPy
• Pandas
• scikit-learn
• faket - for neural style transfer
• IMOD containing system

Optional Dependencies

• CUDA-enabled GPU (recommended for faster style transfer)
• MPI (for parallel processing)

Installation

1. Clone this repository:

git clone https://github.com/ybo-source/fakET_polnet.git

2. Install faket (follow instructions from the) faket repository

3. Activate faket-GPU environment

Directory Structure

Before running the pipeline, set up your directory structure as follows:

base_directory/
├── simulation_dir_<simulation_index>/          # Simulation data (Required-from polnet)
│   └── all_v_czii/           # Simulation name
├── style_tomograms_<style_index>/         # Style tomograms for projection (Required)
├── faket_data/
│   └── style_micrographs_0/   # Projected style micrographs (auto-created)
├── micrograph_directory_0/    # Output directories (auto-created)
├── train_directory_0/         # Training data (auto-created)
└── pipeline.py               # This script

Usage

Basic Usage

python pipeline.py /path/to/your/base_directory

Advanced Usage with Custom Parameters

python pipeline.py /path/to/your/base_directory \
    --micrograph_index 0 \
    --style_index 0 \
    --simulation_index 0 \
    --faket_index 0 \
    --train_dir_index 0 \
    --static_index 0 \
    --tilt_start -60 \
    --tilt_end 60 \
    --tilt_step 3 \
    --detector_snr 0.15 0.20 \
    --simulation_name "all_v_czii" \
    --faket_gpu 0 \
    --faket_iterations 5 \
    --faket_step_size 0.15 \
    --random_faket

Parameters

Required Arguments

• base_dir: Base directory containing simulation and style directories

Index Parameters

• --micrograph_index: Micrograph index (default: 0)
• --style_index: Style index (default: 0)
• --micrograph_directory_index: Micrograph directory index (default: 0)
• --simulation_index: Simulation index (default: 0)
• --faket_index: Faket index (default: 0)
• --train_dir_index: Train directory index (default: 0)
• --static_index: Static index (default: 0)

Tilt Series Parameters

• --tilt_start: Tilt series start angle (default: -60)
• --tilt_end: Tilt series end angle (default: 60)
• --tilt_step: Tilt series step size (default: 3)

Simulation Parameters

• --detector_snr: Detector SNR range (default: [0.15, 0.20])
• --simulation_name: Simulation name (default: "all_v_czii")

Style Transfer Parameters

• --faket_gpu: GPU device ID for faket (default: 0)
• --faket_iterations: Number of iterations for faket style transfer (default: 5)
• --faket_step_size: Step size for faket (default: 0.15)
• --faket_min_scale: Minimum scale for faket (default: 630)
• --faket_end_scale: End scale for faket (default: 630)
• --random_faket: Use random faket style transfer (default: True)
• --denoised: Use denoised style micrographs (default: False)

Pipeline Steps

1. Directory Validation: Checks for required simulation and style directories
2. Style Micrograph Projection: Projects style tomograms to micrographs (if needed)
3. Label Transformation: Processes simulation labels for training
4. Micrograph Projection: Generates clean and noisy micrographs from simulations
5. Style Transfer: Applies neural style transfer using faket
6. Reconstruction: Reconstructs tomograms from style-transferred micrographs
7. Data Organization: Prepares final training dataset structure

Output Structure

After successful execution, the pipeline creates:

base_directory/
├── micrograph_directory_{index}/
│   ├── micrographs_output_dir_{index}/
│   │   ├── Micrographs/          # Projected micrographs
│   │   └── TEM/                  # TEM simulations
│   ├── faket_mics_style_transfer_{index}/  # Style-transferred tomograms
│   └── snr_list_dir/             # SNR metadata
├── style_micrographs_output_{index}/       # Temporary style projection
├── faket_data/
│   └── style_micrographs_{index}/ # Final style micrographs
└── train_directory_{index}/
    └── static_{index}/
        ├── ExperimentRuns_faket/  # Style-transferred training data
        └── ExperimentRuns_basic/  # Basic training data

Customization

Adding New Style Sources

1. Place new style tomograms in base_directory/style_tomograms_{new_index}/
2. Run the pipeline with --style_index {new_index}

Modifying Style Transfer Parameters

Adjust faket parameters for different style transfer effects:

python pipeline.py /path/to/base_directory \
    --faket_iterations 10 \
    --faket_step_size 0.1 \
    --faket_min_scale 500 \
    --faket_end_scale 800

Using Pre-projected Style Micrographs

If you already have style micrographs, place them in: base_directory/faket_data/style_micrographs_{index}/ The pipeline will skip projection and use them directly.

Troubleshooting

Common Issues

1. Missing Dependencies: Ensure faket and svnet utilities are properly installed
2. Directory Not Found: Verify the base directory contains required subdirectories
3. CUDA Errors: Check that CUDA is properly configured and the specified GPU is available
4. Memory Issues: Reduce batch size or use fewer iterations for style transfer

Logging

The pipeline provides detailed logging. Key information includes:

• Directory validation status
• Style transfer progress
• Reconstruction steps
• Output file locations

Contributing

Contributions are welcome! Please feel free to submit pull requests or open issues for bugs and feature requests.

Support

For questions and support, please open an issue on GitHub.