Deep Convolutional Neural Network with Transfer Learning for Detecting Pneumonia on Chest X-Rays

Increasing medical abnormalities has led to life insecurity. Detection and treatment of such abnormalities at an early stage can help save lives. Developing tools and algorithms to make human life better is the main purpose of technology. Detection of such abnormalities through X-ray images is a problem which requires a better solution. There is a need to develop reliable and interpretable deep learning models that can detect such abnormalities with high accuracy. Many models have been developed over time but still there is scope for improvement. In this study, an automatic diagnostic tool based on neural network framework is proposed for the diagnosis of pediatric pneumonia from chest X-ray images. For this task, a deep convolutional neural network model with transfer learning has been proposed. Before passing images to model, preprocessing is done using filtering, gamma correction, equalization, and compression methods. The proposed model is compared with ResNet, ImageNet, Xception, and Inception in terms of precision, recall, accuracy, and ROC accuracy score. Experimentation is done on standard X-ray dataset collected from the Women and Children's Medical Center. The experimental results clearly reveal that our proposed algorithm gives better results than other conventional models.