Automated Pneumonia Detection from Chest X-Rays Using Convolutional Neural Networks: Advancements, Optimization, and Performance Analysis

Pneumonia, an almost fatal bacterial disease generally affecting one or both lungs and brought on by the bacterium Streptococcus pneumonia, can strike any human. The World Health Organisation (WHO) reports that pneumonia accounts for one in three fatalities in India. Chest X-rays are used to detect pneumonia, require the examination of qualified radiotherapists. Establishing an automated approach to identify pneumonia would therefore be advantageous to treat the illness as soon as feasible, particularly in distant places. Convolutional Neural Networks (CNNs) have attracted a lot of interest for disease diagnosis because of the effectiveness of deep learning algorithms in evaluating medical imagery. Higher accuracy, scalability, and automation opportunities are only a few of the benefits of the suggested CNN-based approach over conventional methods. This study proposes to use deep learning methods to find intricate patterns and characteristics from medical imaging data that are suggestive of pneumonia, hence enhancing diagnosis accuracy. Additionally, picture categorization jobs benefit much from features obtained from big datasets using pre-trained CNN models. This study assesses the capability of these models in this work to diagnose abnormal and normal chest X-rays using pre-trained CNN models as feature extractors and various classifiers subsequently. There is an analytical identification of the optimal CNN model for the task.