AI-Based Pneumonia Detection Using a Deep Context Axial Reverse Capsule Neural Network with Sea-Horse Optimizer

Pneumonia detection with AI uses advanced deep learning algorithms for identifying the patterns in the chest X-ray images which makes diagnosis faster and efficient for physicians. AI-powered radiography improved the diagnostic approaches, even then existing methods face some limitations such as reduced accuracy and high computational costs. The Deep Context Axial Reverse Capsule Neural Network with Sea-Horse Optimizer (ADpCARCNNet+SHO) for pneumonia detection is proposed to overcome these challenges. This framework processes chest X-ray images from the Chest X-ray 14 dataset, beginning with Color Wiener Filtering (CWF) for noise reduction. Key features are then extracted using the Dynamic Range Quantized Discrete Haar Wavelet Transform (DRQDHWT), followed by classification using the ADpCARCNNet, optimized with the Sea-Horse Optimizer (SHO) to ensure high precision. Proposed method achieves a remarkable 99.9% accuracy on the Chest X-ray 14 dataset, outperforming existing techniques. The results demonstrate improved diagnostic accuracy, faster image analysis, enhanced reliability, and reduced computational cost, highlighting the potential of the ADpCARCNNet+SHO framework for efficient and reliable AI-driven pneumonia detection.