Deep Learning for X-ray Vision

Original Source Here


TinkerCAD is a free, online 3D modelling software tool created by Autodesk. It features simple to use shapes and hollow object creation, thus making it suitable for creation of pore-like objects for the purpose of simulation.


Meshmixer is a free 3D computer aided design (CAD) modelling software tool. It was used to combine several complex pore STL files together for exporting as a single object to aRTist software for CT scan simulation.

Analytical RT inspection simulation tool (aRTist)

aRTist is a computer simulation program for quantitative description of radiographic testing (radiation source, attenuation of radiation, x-ray film and digital detectors, interactive virtual scene with CAD interface).

VGStudio MAX

Volume Graphics Studio MAX is a software program that was developed to provide the user with a powerful way to process and visualize volumetric data. It was used to reconstruct the 3D image after performing a simulated CT scan using aRTist.

ImageJ (Fiji)

ImageJ is an image processing program written in Java. It features an extensive range of capabilities. This includes scaling, cropping, reduction of noise and saving of image data in multiple formats such as .png, .jpeg, .tif, .tiff. The Weka Trainable Segmentation plugin was also used in this project to perform semi-automatic labelling of the images.

Google Colab Pro

For the training of models, the Google Colab Pro Tesla P100-PC1E-16GB was used. The Google Colab Pro Edition also comes with a faster Intel(R) Xeon(R) CPU @ 2.30GHz as compared to the free version of Intel(R) Xeon(R) CPU @ 2.20GHz.

Weights and Biases

Weights and Biases is a developer tool used to track, compare and visualize machine learning experiments. All experiments carried out in this project were recorded on Weights and Biases.



Trending AI/ML Article Identified & Digested via Granola by Ramsey Elbasheer; a Machine-Driven RSS Bot

%d bloggers like this: