Publication

Abstract : In the realm of medical imaging, the detection of lung pneumonia through automated methods has garnered significant attention due to its potential to augment diagnostic processes. This study proposes a novel approach, termed Federated CNN Ensemble, aiming to enhance the accuracy and robustness of pneumonia detection from chest X -ray images through federated learning and ensemble techniques. The proposed method leverages convolutional neural networks trained locally at distributed medical institutions, thereby preserving data privacy and security while harnessing collectiveintelligence. By aggregating the learned models through an ensemble strategy, our approach achieves superior performance in pneumonia detection compared to standalone CNN models. This research contributes to the advancement of automated medical diagnosis systems, offering a promising avenue for scalable and collaborative disease detection in healthcare settings.