Skip to content

Latest commit

 

History

History
139 lines (98 loc) · 3.84 KB

README.md

File metadata and controls

139 lines (98 loc) · 3.84 KB
Raw

3D CT Scan Viewer

ChatGPT Interface

A graphical user interface (GUI) for viewing 3D CT scans and overlaying ground truth and predicted segmentation masks.


Required Libraries

Install the required libraries using pip:

pip install numpy nibabel PyQt5 matplotlib

How to Use

Here is the updated README to clarify both methods of naming the folders and files:

Prepare Directories

Ensure the following directories exist in the project root:

  • CT
  • Ground_truth
  • Predicted

If they do not exist, the script will create them automatically.

Naming Conventions

You can name your files using one of the following two methods:

  1. Numeric Only IDs:

    • CT scans: Place files in the CT directory. Name them as <subject_id>.nii.gz.

      Example:

      CT/
├── 001.nii.gz
├── 002.nii.gz
└── 003.nii.gz

    • Ground truth segmentations: Place files in the Ground_truth directory. Name them as <subject_id>.nii.gz.

      Example:

      Ground_truth/
├── 001.nii.gz
├── 002.nii.gz
└── 003.nii.gz

    • Predicted segmentations: Place files in the Predicted directory. Name them as <subject_id>.nii.gz.

      Example:

      Predicted/
├── 001.nii.gz
├── 002.nii.gz
└── 003.nii.gz

  2. Alphanumeric IDs:

    • CT scans: Place files in the CT directory. Name them as <subject_id>.nii.gz.

      Example:

      CT/
├── SUB_001.nii.gz
├── SUB_002.nii.gz
└── SUB_003.nii.gz

    • Ground truth segmentations: Place files in the Ground_truth directory. Name them as <subject_id>.nii.gz.

      Example:

      Ground_truth/
├── SUB_001.nii.gz
├── SUB_002.nii.gz
└── SUB_003.nii.gz

    • Predicted segmentations: Place files in the Predicted directory. Name them as <subject_id>.nii.gz.

      Example:

      Predicted/
├── SUB_001.nii.gz
├── SUB_002.nii.gz
└── SUB_003.nii.gz

Execute the script to start the GUI:

python viewer.py

Using the GUI

  • Use the left < and right > arrow buttons to navigate through different subjects. You can use either numeric or alphanumeric IDs for the subjects (e.g., 001, SUB_001).
  • Enter the subject ID directly in the input field to load a specific subject.
  • View Controls: Switch between axial, coronal, and sagittal views.
  • Toggle contour display and adjust contour line width.
  • Adjust colors for labels and opacity of overlays.
  • Adjust the slider to navigate through slices.
  • Change the minimum and maximum intensity values to adjust the CT scan contrast.
  • Use the ▶ button to animate through the slices automatically.
  • Toggle the overlay of ground truth and predicted segmentation masks using the Toggle Overlay button.
  • Plot intensity histogram

Support Us

If you find it helpful, consider supporting us in the following ways:

License

This project is licensed under the MIT License. See the LICENSE file for details.

Contact

For any inquiries, please contact AITwinMinds@gmail.com.