/3d_localization_with_cai

This repository contains both the code and the image data for the IOP Physics in Medicine and Biology publication "3D-Localization of Single Point-Like Gamma Sources with a Coded Aperture Camera"

Primary LanguagePythonGNU General Public License v3.0GPL-3.0

DOI

3d_localization_with_cai

This repository contains both the code and the image data for the IOP Physics in Medicine and Biology publication "3D-Localization of Single Point-Like Gamma Sources with a Coded Aperture Camera"

Installation

The scripts were implemented and executed with Python 3.8.18. Further, the following packages are required:

  • NumPy (1.24.4)
  • tensorflow (2.10.1)
  • SciPy (1.10.1)
  • vtk (9.3.0)
  • matplotlib (3.7.2),
  • Pandas (2.0.3)
  • Seaborn (0.12.2)
  • openCV (4.6.0)
  • bottleneck (1.3.5)

Decompress the provided zip files, so that all tiff files are in ./3d_reconstructions/experimental/ and ./3d_reconstructions/mc_simulation/ accordingly.

Running the Localization Methods

Running the script "analysis_and_plots.py" collects all the 3D reconstructions and performs successively different analysis. Each time a Pandas dataframe is produced and automatically copied into the clipboard. If that causes problems the corresponding line containing "resfex.round(2).to_clipboard()" can be commented out or deleted. In addition to the dataframe for each analysis a grouped boxplot diagramm is generated and displayed.

The localization methods Iterative Source Localization (ISL) and the Center Of Mass (COM) methods are implemented in the script "localize.py".