Using Python to classify DICOM computerized tomography images. Uses the TensorFlow framework. Click here to view this project's website. To install dependencies, make sure that pip is running on your system.
By using this software, you agree to the terms and conditions outlined in the License. This results generated by this software should not be taken as medical advice and should not be a substitute for seeing a medical professional. Furthermore, the outputs of this software should not be considered conclusive unless verified by a medical professional.
Google's open-source machine learning library. DeepCT uses the TensorFlow framework.
For the software to be converted to an executable file, some newer versions are not compatible. The latest version tested with DeepCT is version 1.10, which is incompatible with Python versions 3.7 and newer.
Note: The GPU version of TensorFlow has not been used with this software.
pip install tensorflow==1.10
To check if TensorFlow is running on your system, run:
import tensorflow as tf
print(tf.__version__)
If installed correctly, it should output version 1.10.
Used to plot the data, visually display the data, and create .png images.
pip install matplotlib
Used to convert the DICOM images into .png format.
pip install pydicom
A fundamental library for machine learning.
pip install scipy
The foundation for Tensorflow and SciPy. Installed with TensorFlow and SciPy by default.
pip install numpy
To check if NumPy is running on your system, run:
import numpy as np
print(np.__version__)
Used to work with the created .png images (originally in .dcm format).
pip install pillow
Installed with TensorFlow by default. Used to save and load existing ML models.
pip install h5py
Used to convert the .py file(s) into an executable file (Windows).
pip install pyinstaller