Alien clustering for BCDI data

Introduction

This package use clustering algorithms to find aliens in BCDI data with minimal users input with the following steps :

• preprocessing : filter and rescale BCDI data in custom log scale
• mask creation : intensity threshold mask and possible smoothing
• clustering : clustering using sklearn DBSCAN algorithm
• filtering : filter out clusters
• user cluster selection : widget selection of the alien cluster from the user
• alien mask creation : create the final alien mask and save result

Files description

• plot_utiltities.py : Some general plotting functions
• alien_removal_3D_utilities.py : All functions used for the alien mask creation
• Alien_removal_notebook.ipynb : jupyter notebook with a step-by-step procedure to make the alien mask
• example_data : folder with BCDi data as a tutorial
• requirements.txt : file containing required python module versions (ipywidgets CheckBox can be broken in recent versions)

Get started

• clone the repository

  git clone https://github.com/ewbellec/alienclustering.git

• open the jupyter notebook Alien_removal_notebook.ipynb

• follow the instructions

• If you encounter issues using this code, email ewen.bellec@esrf.fr

Video tutorial

alien_clustering_tutorial.mp4

Illustrations