Combining Deep Learning Architectures with Fuzzy Logic for Robust Pneumonia Detection in Chest X-Rays

Early and accurate detection of pneumonia from chest X-ray images is essential for improving treatment and clinical outcomes. However, medical imaging datasets often exhibit class imbalance and considerable uncertainty in the extracted image features, which complicates the use of conventional classification methods and motivates the adoption of advanced techniques such as deep learning and fuzzy logic. This study proposes a hybrid approach combining deep learning architectures (VGG16, EfficientNetV2, MobileNetV2, ResNet50) for feature extraction with fuzzy logic-based classifiers, including models that handle uncertainty and fuzzy rules. Additionally, feature selection techniques were employed to enhance feature extraction. The model achieved an overall accuracy of approximately 93.5%. For the Abnormal class, precision was around 97.4%, sensitivity was 95.4%, and the F1-score was approximately 96.4%. For the Normal class, precision was near 88.5%, sensitivity was 89.3%, and the F1-score was about 88.9%. The results demonstrate that integrating deep feature extraction with fuzzy logic improves diagnostic accuracy and robustness, providing a reliable tool for clinical support. Future research will focus on optimizing model efficiency and interpretability.