This repository contains code for generating a model to detect pneumothorax from user-uploaded images using the ResNet-50 V2 architecture. The project focuses on binary image classification, specifically distinguishing between images with pneumothorax and images without pneumothorax. The repository includes several essential steps to train and evaluate the model.
The dataset used in this project can be downloaded from this link. It provides a collection of images for both classes: with pneumothorax and without pneumothorax.
-
imagesplitting.py: This step involves splitting the dataset images into training and validation sets. It ensures proper allocation of data for training and evaluation purposes. -
imageaugmentation.py: The augmentation step enhances the dataset by applying transformations to existing images. It helps improve the model's ability to generalize and handle variations in input images. -
train_val_split.py: This step splits the augmented dataset into training and validation subsets. It enables model training on the training set and validation for evaluating model performance. -
model.py: The model creation step involves defining and building the binary classification model architecture. It utilizes the ResNet-50 V2 pre-trained model as the base architecture for detecting pneumothorax. -
app.py: The Streamlit application allows users to upload their own images and utilizes the trained model to predict if the image contains pneumothorax or not.
To use this repository and detect pneumothorax from images, follow these steps:
-
Download the dataset from the provided link and place it in a folder named
dataset. -
Execute the scripts in the given order:
imagesplitting.py,imageaugmentation.py,train_val_split.py,model.py. -
Once the model is trained, run the Streamlit application using
app.py. It will open a web interface where you can upload your own images. -
Upload an image through the application and click the "Predict" button. The model will analyze the image and provide a prediction if the image contains pneumothorax or not.
Feel free to explore the code, modify it, and experiment with different model architectures or techniques to improve the detection accuracy.
Contributions to this repository are highly welcome. If you have any ideas, suggestions, or improvements, please feel free to fork the repository and submit a pull request. Let's collaborate to enhance the pneumothorax detection model!
This project is licensed under the MIT License. You are free to use and modify the code for your own purposes.