Stripe Begone

Stripe Begone is a GUI-based tool designed to remove stripe artifacts from microscopy images (specifically SEM/FIB-SEM) using Fourier space masking and advanced reconstruction algorithms.

Features

• Interactive Fourier Masking: Define wedges to mask specific frequency components corresponding to stripes.
• Automated Angle Detection: Uses Radon transform variance analysis to automatically detect stripe orientation.
• Advanced Reconstruction Methods:
  • FISTA (Fast Iterative Shrinkage-Thresholding Algorithm)
  • POCS (Projection Onto Convex Sets)
  • Total Variation (TV) Denoising
  • Weighted L1 Minimization
  • KNN Imputation
  • Zero Fill
• Texture Recovery: Restore non-stripe high-frequency details lost during reconstruction.
• Contrast Enhancement: Built-in tools for histogram adjustment.
• GPU Acceleration: Utilizes PyTorch for accelerated reconstruction if CUDA is available.

Requirements

• Python 3.8+
• PySide6
• NumPy
• SciPy
• scikit-image
• scikit-learn
• Matplotlib
• PyTorch

Installation

Prerequisites

• Python 3.14 or higher recommended (Though it is not tested it should run with earlier version of Python 3).
• uv (optional, but recommended for building).

Installation

Quick Start with uv

The fastest way to try pyTRACTnmr without installation is:

uvx IteraStripe

Using pip

pip install IteraStripe

From Source

1. Clone the repository:

   git clone https://github.com/debadutta-patra/IteraStripe
   cd pyTRACTnmr
   

2. Install the package:

   Using uv (fastest):

   uv pip install .
   

   Using standard pip:

   pip install .
   

Usage

1. Launch the Application:

   python src/IteraStripe/main.py
   

2. Load an Image:

   • Click File -> Open Image or drag and drop an image file onto the window.

3. Configure Masks:

   • Use the Wedge Parameters controls to define the stripe angle and width.
   • Click Auto to attempt automatic angle detection.
   • Click Add Wedge to register the mask.

4. Remove Stripes:

   • Select a method from the Processing panel (Default: FISTA Reconstruction).
   • Click Remove Stripes.

5. Refine and Save:

   • Use the tabs to view the Reconstruction, Difference map, or FFT.
   • Apply Texture Recovery if the result is too smooth.
   • Save the final image.