This project aims to develop deep learning models for the classification of chest X-ray images into three categories: Viral Pneumonia, COVID-19, and Normal. The models are trained and evaluated using a dataset containing images from these three classes.
- Click on the Dataset to download the zip Dataset file.
- The dataset used in this project is the COVID-19 Chest X-ray dataset, which consists of X-ray images collected from various sources. The dataset is divided into training and testing sets, ensuring a balanced distribution of images across the three classes.
Three different models are implemented for the classification task:
-
CNN Model: This model is designed specifically for this project and consists of multiple convolutional layers, max-pooling layers, and fully connected layers.
-
VGG16: The VGG16 model, a popular pre-trained architecture, is used as a base model and fine-tuned for this classification task. Additional layers are added on top to adapt it to the specific problem.
-
ResNet50: The ResNet50 model, another pre-trained architecture, is also used as a base model. It employs residual blocks and skip connections to enable better learning of complex features.
To get started with the project, follow these steps:
-
Clone the repository:
git clone https://github.com/Pavan042/Covid_19_Image_Classification_CNN_Capstone_Project.git -
Preprocess the dataset: Resize the images, normalize the pixel values, and split the data into training and testing sets.
-
Train the models: Run the training script for each model to train them on the training set.
-
Evaluate the models: Use the trained models to predict the labels of the test set and evaluate their performance using metrics such as accuracy, precision, recall, and F1-score.
The models are evaluated based on their performance metrics, including accuracy, precision, recall, and F1-score. The results show that the ResNet50 model achieves the highest accuracy and performs the best in classifying COVID-19 chest X-ray images.
To further improve the models and their application, consider the following:
- Experiment with other pre-trained models and architectures to explore their performance and suitability for the classification task.
- Augment the dataset with additional labeled examples or consider using transfer learning techniques to leverage larger external datasets.
- Explore ensemble methods by combining the predictions of multiple models to enhance classification accuracy.
- Deploy the best-performing model as a web or mobile application for real-time COVID-19 detection.
In conclusion, this project demonstrates the development and evaluation of deep learning models for COVID-19 chest X-ray classification. The models achieve high accuracy and provide valuable insights for the detection and diagnosis of COVID-19 cases. By leveraging pre-trained architectures and fine-tuning them for the specific task, accurate classification results can be obtained. The project highlights the potential of deep learning in aiding healthcare professionals in identifying COVID-19 cases from chest X-ray images.