This initiative endeavors to design and deploy a Convolutional Neural Network (CNN) model dedicated to identifying COVID-19 in chest X-ray images. The primary objective is to construct a highly accurate and efficient model, aiming to make a substantial impact on the prompt and precise diagnosis of COVID-19. This contribution is anticipated to enhance patient care and mitigate the spread of the disease through effective management strategies.
To fulfill the project objective, we propose the training of a Convolutional Neural Network (CNN) model using a comprehensive dataset of chest X-ray images that includes both COVID-19-positive and negative cases. The primary aim is to establish a reliable and efficient method for the detection of COVID-19. Once trained, the model will possess the capability to classify new chest X-ray images as either COVID-19-positive or negative. This classification functionality is intended to assist healthcare professionals in making timely and well-informed decisions, thereby enhancing the overall diagnostic process for COVID-19.
Chest X-rays offer a quicker diagnostic turnaround, taking only minutes per patient, making them ideal for high-throughput settings. CT scans of the thorax, while providing high-resolution images, are chosen for detailed lung visualization, aiding in early detection of COVID-19 manifestations.
For our COVID-19 detection project, we utilized the "Chest X-Ray for COVID-19 Detection" dataset from Kaggle, featuring diverse chest X-ray images encompassing both COVID-19-positive and negative cases. This publicly available dataset, comprising images sourced from multiple medical institutions, is stored in jpeg/jpg/png format. The dataset, organized within the "Dataset" directory, is further categorized into "Prediction," "Train," and "Val" (Validation) subdirectories, collectively containing hundreds of chest X-ray images for comprehensive analysis. Link-to-dataset