Pneumonia Detection Using Deep CNN Algorithms

The detection of pneumonia using medical imaging is critical for prompt and effective treatment. This study looks at how deep convolutional neural networks (CNNs) can be used to detect pneumonia using a CT scan dataset from Kaggle. The data set is divided into three categories pneumonia COVID-19 and normal. For this study the Pneumonia and COVID-19 classes were combined into a single ‘Disease’ category. The proposed model designed to address challenges such as high variances in infection characteristics and the scarcity of annotated data which is a combination of ResNet152V2 and CNN networks were then used for classification leveraging their powerful feature extraction capabilities achieves significant results. For the datasets mentioned above the suggested model yields validation accuracy of 90.56% and accuracy of 99.09%.