/ct_lung_segmentation

Robust segmentation of lung and airway in CT scans

Primary LanguageJupyter NotebookMIT LicenseMIT

CT Lung Segmentation

This repository contains the codes for lung and airway segmentation from CT scans, used in "Discriminative Localization in CNNs for Weakly-Supervised Segmentation of Pulmonary Nodules", MICCAI'17.

Files

  • demo_ct_lung_segment.ipynb: demo code to walk through the steps implementated in main.py.
  • main.py: main code of the segmentation.
  • segment_lung.py: coarse segmentation of lung & airway.
  • segment_airway.py: extraction of airway.
  • utils.py: utility functions.
  • data: input sample.
  • result: output sample.

Note: the sample image is down-sampled from a CT scans in the LIDC-IDRI dataset.

Requirements

  • numpy
  • nibabel
  • scipy
  • scikit-image

Usage

    $ git clone https://github.com/wanwanbeen/ct_lung_segmentation
    $ cd ct_lung_segmentation
    $ python main.py