COVID-19 Detection using Deep Learning

Overview

This project focuses on detecting COVID-19 from chest X-ray images using deep learning techniques. The goal is to classify images into COVID-positive and COVID-negative categories using a Convolutional Neural Network (CNN).

Dataset

• Source: Publicly available datasets of chest X-ray images from Kaggle and other medical repositories.
• Preprocessing: Image resizing, normalization, and augmentation to improve model robustness.

Methodology

1. Data Preprocessing:

   • Loaded and preprocessed images using OpenCV and TensorFlow.
   • Augmented data to prevent overfitting and improve generalization.

2. Model Architecture:

   • Implemented a CNN model using TensorFlow/Keras.
   • Used transfer learning with pre-trained models like VGG16 and ResNet50 for better accuracy.

3. Training & Evaluation:

   • Split data into training and validation sets.
   • Optimized the model using Adam optimizer and categorical cross-entropy loss.
   • Evaluated model performance using accuracy, precision, recall, and F1-score.

Results

• Achieved 90%+ accuracy on the validation dataset.
• Improved classification performance using transfer learning.
• Developed visualizations of Grad-CAM to interpret model decisions.

Technologies Used

• Programming Languages: Python
• Frameworks & Libraries: TensorFlow, Keras, OpenCV, Matplotlib, NumPy, Pandas
• Model Training: GPU-accelerated training using TensorFlow
• Evaluation Metrics: Accuracy, Precision, Recall, F1-score

Future Improvements

• Expand the dataset for better generalization.
• Implement ensemble learning for improved performance.
• Deploy the model as a web application using Flask or FastAPI.