Sequential Transfer Learning Model for Pneumonia Detection Using Chest X-ray Images
Aditya Kumar, Leema Nelson, S. Gomathi
Abstract
This work utilises a sequential convolutional neural network (CNN) model for the detection of Pneumonia in chest X-ray images, focusing on improving diagnostic accuracy. The model utilises transfer learning, utilising a pre-trained neural network architecture to achieve the highest accuracy rate of 92.15% on the test dataset. Moreover, this represents a significant advancement over prior methodologies, which attained an accuracy of 90.22%. The robust precision, recall, and F1-score metrics highlighted the model’s performance, affirming its proficiency in distinguishing between pneumonia (a medical condition) and normal (a healthy state) cases. This research contributes to medical diagnostics, offering a promising solution for healthcare professionals in their quest for improved pneumonia detection from chest X-ray images. Moreover, the work identifies potential areas for further refinement, encompassing hyperparameter optimisation, improved interpretability techniques, and practical deployment, all aimed at elevating the utility of this transfer learning model in medical diagnostics and overall healthcare. This research signifies a notable significant advancement in utilising the power of deep learning for medical disease detection.