Skip to content

HelenaMaslowska/Tomograph

Repository files navigation

Tomograph experiment

Hi! We created a project which explains how to convert image into a sinogram and vice versa with filtering, normalization and visualization. You can see the progress by moving the progress bar and saving to a DICOM file! Here is a main code, enjoy!

Used libraries

  • numpy - filtering with convolve arrays and math
  • scikit-image - draw lines using ski.draw.line_nd(E, D), image
  • matplotlib - plots
  • pydicom - extension of a file
  • cv2 - filtering with cv2.filter2D
  • ipywidgets - interact (slider)

Slider

One of the most interesting part of an experiment. Move progress bar to check how is Radon transform actually works! I wish I don't spoiler too much...

image

Summary

All the result of an experiment is summarized here by MSE measure at the bottom of the page.

Source image

image

Sinogram created by Radon transform

image

Sinogram from OpenCV library (cv2.filter2D)

image

Sinogram with numpy library (np.convolve)

image

Invert Radon transform

image

Invert Radon transform with openCV

image

Invert Radon transform with numpy

image

Authors

  • Marcin Szwarc
  • Helena Masłowska