CT-Lung-Segmentation

This repository contains a Pytorch implementation of Lung CT image segmentation Using U-net

Figure 1: Original CT images

Figure 2: Ground-truth Segmentation Mask

Figure 3: Predicted Segmentation Mask

Requirement

jupyter

numpy

opencv-python

Pillow

torch

torchvision

tqdm

pip install -r requirements.txt

Dataset

1. Download the data from Kaggle/Finding and Measuring Lungs in CT Data
2. With totally 267 CT slices, I randomly select 200 slices for training and 67 for testing