Detection of Lung Diseases Using Deep Transfer Learning-Based Convolution Neural Networks

Human lungs are susceptible to numerous diseases, emphasizing the importance of early detection and timely intervention to minimize complications. This study explores the transformative potential of Artificial Intelligence (AI) through transfer learning and Convolutional neural networks (CNNs) in predicting these lung diseases through meticulous analysis of chest X-ray images. In this paper, we have compared the detection performance of four variants of Convolutional neural networks on a benchmark dataset. Our study focuses on the comparative evaluation of existing deep learning models, including Densenet121, ResNet18, GoogLeNet, and MobileNet, all of which have been pre-trained on large image datasets to classify 4-class pathologies in chest X-ray images. With their foundations in transfer learning, these models have shown promising performance for lung Disease detection. Among these four models, the Detection performance of MobileNet and ResNet18 is quite encouraging compared to DenseNet121 and GoogLeNet. This approach could revolutionize the early detection and treatment of lung diseases, thereby enhancing patient outcomes and healthcare efficiency by providing insights into the strengths and weaknesses of these existing models.