Overview of fine-tuning CNN-Based Models for X-ray Image Classification

A lung infection is usually the cause of pneumonia, a common medical condition. It irritates the lungs' tissues and reduces their functionality. The severity of pneumonia can vary from a minor illness to a serious one. Identifying the exact infection causing the problem can be difficult. Diagnosis is often based on symptoms and physical examination, sometimes with chest X-rays. On the other hand, reviewing chest X-rays is a challenging and subjective task. In this work, we improve the CNN architecture to improve the X-ray image classification score performance. The objective of this study is to evaluate the fine-tuning of ResNet50V2. The ensemble technique that has been recommended yields very strong classification results, outperforming other models with an improvement of almost 97% in accuracy.