Skip to content

Latest commit

 

History

History
68 lines (54 loc) · 2.68 KB

README.md

File metadata and controls

68 lines (54 loc) · 2.68 KB
Raw

Lung-Segmentation

This repository contains code and resources for segmenting abnormal regions in chest X-ray images using deep learning techniques. The goal is to accurately identify pneumonia, tumors, or other lung-related abnormalities.

Dataset

  • The dataset used in this project is sourced from Kaggle and is named nikhilpandey360/chest-xray-masks-and-labels.
  • It includes original chest X-ray images along with corresponding lung masks.
  • Ensure you have downloaded the dataset before proceeding.

Data Preparation

  1. Data Augmentation:

    • Apply data augmentation techniques (e.g., flipping, rotation) to increase dataset diversity.
    • Augmented data helps improve model generalization.

  2. Preprocessing:

    • Crop or resize images to focus on the lung area.
    • Normalize pixel values to a consistent range (e.g., [0, 1]).

Model Architecture

  • We experimented with two residual ResNets:
    1. Conventional ResNet:
      • Utilizes standard CNN layers.
      • Captures hierarchical features through residual connections.
    2. Depthwise Separable ResNet:
      • Employs depthwise separable convolutions for efficiency.
      • Reduces the number of parameters while maintaining performance.

Performance Evaluation

  • We assessed model performance using the Dice coefficient (F1 score).
  • This metric quantifies the overlap between predicted and true positive regions.

Visualization

  • Created segmentation figures to visualize model predictions.
  • Compare predicted masks with actual abnormalities in X-rays.

Next Steps

  • Fine-tune hyperparameters, explore additional architectures, or incorporate other evaluation metrics.
  • Continue refining the model based on performance and domain-specific requirements.

Instructions for Kaggle API

  1. Download Kaggle API:

    • Install the Kaggle API by running pip install kaggle.

  2. Kaggle API Token:

    • Go to your Kaggle account settings and generate an API token.
    • Save the token as kaggle.json in the root directory of this repository.

  3. Download Dataset:

    • Use the Kaggle API to download the dataset: 
      kaggle datasets download -d nikhilpandey360/chest-xray-masks-and-labels

  4. Upload Kaggle API Token to Colab/Notebook:

    • If using Colab or Jupyter Notebook, upload the kaggle.json token to your environment.
    • Use the following code snippet: 
      from google.colab import files
files.upload()

  5. Unzip Dataset:

    • Unzip the downloaded dataset: 
      unzip chest-xray-masks-and-labels.zip

Feel free to explore, contribute, and enhance this project! 🌟👩‍⚕️🔍