Tomograph Simulator

Hey there! :]

This little app is a basic simulator for how a CT (Computed Tomography) scanner works. You feed it an image (like a photo or even a medical DICOM file), and it pretends to scan it.

It creates sinogram (which is like the raw data the scanner gets) and then tries to reconstruct the original image back from that sinogram.

GUI - streamlit

IMPORTANT NOTE!

• No filters are applied after Inverse Radon Transformation

Getting it Running

0. Git Clone: Clone Repo! ;}

   git clone https://github.com/CosmoVoxel/tomograf.git

1. Install the Goods: Open your terminal or command prompt and run this:

   pip install -r requirements.txt

2. Run the App: Navigate to the folder where you saved app.py in your terminal and run:

   streamlit run app.py

   Your web browser should pop open with the app!

Links:

• https://en.wikipedia.org/wiki/Tomographic_reconstruction
• https://www.mathworks.com/help/images/radon-transform.html
• https://www.mathworks.com/help/images/the-inverse-radon-transformation.html

How to Use It

• Upload Image: Use the sidebar on the left to upload an image file. It can handle common types like JPG, PNG, and also DICOM (.dcm) files. If you upload a DICOM, it'll try to grab patient info from it.
• Tweak Parameters:
  • Angular Step: How much the pretend scanner rotates between each 'shot'. Smaller angles mean more projections (usually better quality, but slower).
  • Number of detectors: How many sensors collect data for each projection.
  • Detector spacing: How far apart the sensors are.
• See the steps?: Check the Show intermediate steps box if you want to watch the sinogram and reconstruction build up step-by-step. It's neat but takes longer.
• Patient Info: Fill this stuff in if you want it saved in the final DICOM file you download later.
• Buttons:
  1. Click Create Sinogram first. This generates the projection data from your image based on the parameters you set.
  2. Once the sinogram is ready, click Reconstruct Image. This uses the sinogram data to build the final image.
• Download: If the reconstruction looks okay you'll see a Save as DICOM button appear below it. Click that to save your result.

Cool Stuff It Does

• Loads JPG, PNG, and DICOM images.
• Tries to read patient info from DICOMs.
• Shows animated step-by-step processing (optional).
• Saves the final reconstructed image as a DICOM file.

What You Need (Requirements)

• Python (Version 3.x recommended)