Hybrid deep learning and machine learning framework for automated pneumonia detection in chest X-ray images

Pneumonia is a dangerous respiratory illness that has to be precisely and promptly diagnosed in order to be treated effectively and prevent consequences. In order to distinguish between pneumonic and normal chest X-ray pictures, a hybrid deep learning technique is proposed in this study. For efficient and complementary feature extraction, the proposed system leverages the strengths of two popular convolutional neural networks, VGG16 and ResNet. Before training the model, the image is enhanced and the lungs are segmented by performing histogram equalisation, normalising contrast, and converting to grayscale. A richer feature representation of input photos is produced by fusing the features of VGG16 and ResNet. A model for identifying pneumonia is classified using the fused feature set. The system processes X-rays of new patients in order to extract features and categorise them using Random Forest (RF) and Support Vector Machine (SVM) classifiers. To increase accuracy and efficiency, feature dimensions are optimised using Principal Component Analysis (PCA). Key Contributions: 1. Dual-CNN feature fusion (VGG16 + ResNet) instead of single-model learning 2. PCA-based dimensionality optimization retaining 95% variance 3. Use of SVM and Random Forest for more interpretable diagnosis instead of CNN softmax.