tomograph - paramterized computer tomography simulator

tomograph is a parametrized computer tomography simulator, which can be run as a standalone script or included as a module in your Python programs. It implements both Radon and reverse Radon transforms. It currently models two types of beam projections: cone and parallel.

Dependencies

Currently tomograph requires:

• Click (click)
• tqdm (tqdm)
• NumPy (numpy)
• Image Processing SciKit (scikit-image)

Their respective required versions can be seen in requirements.txt file.

Installation

Simply run:

pip install tomograph

Tomograph supports Python 3.5, Python 3.6 and Python 3.7.

Usage

tomograph can be used directly through command-line interface or as a Python module:

Commandline

Using command-line interface:

tomograph --model cone --detectors 256 --angle 150 --rotations 512 image.jpg

All command-line arguments can be seen by running:

tomograph --help

Module

Example script that does Radon transform using cone beam model:

from skimage.io import imsave
from tomograph.model import ConeTomograph

# variables
path = 'image.jpg'
detectors = 256
angle = 180
rotations = 512

# one scan per rotation
step = 360 / rotations

# perform Radon transform
tomograph = ConeTomograph(path, detectors, angle)
sinogram = []
for i in range(rotations):
    tomograph.rotate(step)  # rotate tomograph
    scan = tomograph.scan() # do the scan
    sinogram.append(scan)

imsave('radon.bmp', np.array(sinogram).T)

Live demonstration

Beam lines	Radon transform	reverse Radon transform