Radon Transform

A rust and python implementation of the discretized version of the radon transform that I did out of boredom. The main mathematical transformation is written in Rust. The Rust functions are then wrapped up as Python functions using PyO3, which is later used along with other libs like numpy and PIL.

Building the project

To the run the script, you will first have to build the project. You can build this rust project using cargo:

cargo build --release

After building this, navigate to target/release, where you will find the extension randon_lib.dll (if you are on windows) or radon_lib.so (on OSX). Copy this file into the python folder and rename the extension to radon_lib.pyd

Running the project

To call the underlying rust functions, import the radon_lib.pyd that you created earlier in your python file, radon_trial.py for example.

import radon_lib

You must make sure that radon_lib.pyd is present in the same directory as randon_trial.py.

Once this step is done, you can load the image of choice and visualize the Radon and Inverse Radon Transform.

python3 radon_trial.py

Why is there a struct named "Tissue" in the Rust implementation?

Radon Transform is the heart of the Computed Tomography (CT Scan), which is used for non-invasive diagnosis. The CT Scan is capable of reconstructing the 3-D tissue (along with their attenuation coefficient, meaning how dense is the tissue) by taking multiple 2-D X-Ray slices and stiching them up. The 3-D reconstruction of the original tissue is done by a process called "backprojection" aka the Inverse Radon Transform. I found this interesting and thought I would try it, therefore calling the image struct as "Tissue". Feel free to change that name if you want to use this code for some reason.