Low Contrast Detectability for CT (LCD-CT) Toolbox provides a common interface to evaluate the low contrast detectability (LCD) performance of advanced nonlinear CT image reconstruction and denoising algorithms. The toolbox uses model observers (MO) to evaluate the LCD of targets with known locations in test images obtained with the MITA-LCD phantom. The model observer detection accuracy is measured by the area under the receiver operating characteristic curve (AUC) and the detectability signal-to-noise ratio (d’_{snr}). The LCD-CT toolbox can be used by CT developers to perform initial evaluation on image quality improvement or dose reduction potential of their reconstruction and denoising algorithms.
- Regulatory Science Tool: Check the FDA website for a description of the LCD-CT toolbox in the Regulatory Science Tool Catalog
- Creating digital replica of the background and signal modules of the MITA-LCD phantom.
- Simuating sinogram and generate fan-beam CT scans of the digital phantoms based on the publicly available Michigan Image Reconstruction Tolbox (MIRT).
- Estimating low contrast detectability performance from the MITA-LCD phantom CT images using channelized Hoteling model observer with Laguerre-Gauss (LG) channels and two options of Difference-of-Gaussian (DOG) channels and Gabor channels.
Requirements
- Matlab (version > R2016a) or Octave (version > 4.4)
- If the above Matlab or Octave requirements are not met, then conda is required to install Octave using the installation instructions.
If required versions of Matlab or Octave are not available on your system (see how to get matlab version or octave version) then see installation for how to setup an Octave environment to run LCD-CT.
Installation
- Git clone the LCD-CT Toolbox repository:
git clone https://github.com/DIDSR/LCD_CT
cd LCD_CT- *If neither Matlab or Octave are installed or do not meet the version requirements, you can source install.sh to prepare a conda environment. Note: this can take about 10 minutes to complete.
source install.shExpected run time: 10-30 min
- Test the installation
- From the bash command line octave test.m or matlab -batch test.m
- From the Matlab or Octave interactive prompt
>> testExpected run time (Octave): 1 min 30 s
After installing review the LCD-CT Toolkit Documentation and explore the demos to learn how to use the tool to assess low contrast detectability:
- demo_01_singlerecon_LCD.m Expected run time (Octave): 6 s
- demo_02_tworecon_LCD.m Expected run time (Octave): 14 s
- demo_03_tworecon_dosecurve_LCD.m Expected run time (Octave): 19 s
Additional demos of tool usage can be found in additional_demos.
Accessing the Large Dataset
By default the supplied demos and test scripts will use the small dataset (155 images, 16 MB) included in the repository. The boolean use_large_dataset is set to False by default
Line 6 in aa9a907
Changing the variable to true followed by rerunning any of the scripts will download from Zenodo the large dataset (1.3 GB) and use it in subsequent analyses
The following AUC-vs-dose curves were generated by demo_03_tworecon_dosecurve_LCD.m using the large data set available at Zenodo and the LG channelized Hoteling model observer.
>> use_large_dataset = true
>> demo_03_tworecon_dosecurve_LCD
The LCD Phantom Creation code uses functions from Michigan Image Reconstruction Toolkit (MIRT). It should be automatically downloaded and installed when 'demo_test_phantomcreation.m' is run. If the automatic download does not work (this can happen when the matlab/octave unzip() function does not successfully extract all the files), this can be done manually:
- download MIRT from https://github.com/JeffFessler/mirt;
- Unzip MIRT to a local directory;
- In Matlab, Run the file "setup.m" in the MIRT local directory to add all the MIRT subdirectories to the MATLAB workspace;
To test whether the setup is successful, run demo_test_phantomcreation.m.
Issue Tracker | Source Code | Contributing Guide
- How to cite
(for paper) Vaishnav, J.Y., Jung, W.C., Popescu, L.M., Zeng, R. and Myers, K.J. (2014), Objective assessment of image quality and dose reduction in CT iterative reconstruction. Med. Phys., 41: 071904. https://doi.org/10.1118/1.4881148
(for tool) LCD-CT: Low-contrast detectability (LCD) test for assessing advanced nonlinear CT image reconstruction and denoising methods. https://www.fda.gov/medical-devices/science-and-research-medical-devices/lcd-ct-low-contrast-detectability-lcd-test-assessing-advanced-nonlinear-ct-image-reconstruction-and
(for data) Nelson, B., Zeng, R., CT Simulations of MITA Low Contrast Detectability Phantom for Model Observer Assessments. Published online June 1st, 2023. doi:10.5281/zenodo.7996580
If you are having issues, please let us know.
Toolbox developers: Brandon Nelson ([email protected]), PhD, Rongping Zeng, PhD ([email protected])
This software and documentation (the "Software") were developed at the Food and Drug Administration (FDA) by employees of the Federal Government in the course of their official duties. Pursuant to Title 17, Section 105 of the United States Code, this work is not subject to copyright protection and is in the public domain. Permission is hereby granted, free of charge, to any person obtaining a copy of the Software, to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, or sell copies of the Software or derivatives, and to permit persons to whom the Software is furnished to do so. FDA assumes no responsibility whatsoever for use by other parties of the Software, its source code, documentation or compiled executables, and makes no guarantees, expressed or implied, about its quality, reliability, or any other characteristic. Further, use of this code in no way implies endorsement by the FDA or confers any advantage in regulatory decisions. Although this software can be redistributed and/or modified freely, we ask that any derivative works bear some notice that they are derived from it, and any modified versions bear some notice that they have been modified.
The project is licensed under `Creative Commons Zero v1.0 Universal LICENSE`_.
- DIDSR/IQModelo: Statistical Software for Task-Based Image Quality Assessment with Model (or Human) Observers
- DIDSR/VICTRE_MO
- Example of CT image noise insertion code