Globally lung illness is widespread due to changes in the environment, weather, daily life, and other factors. Thus the impact of this disease on physical condition is increasing quickly. It motivates to the development of Machine Learning (ML) and Deep Learning (DL). It can give medical professionals and other researchers guidance on how to use DL to detect lung disease. In this paper, a novel Elite Opposition Learning Based Transfer Residual Neural Network (EOL-TRNN) is introduced for the identification of pneumonia from a Chest X-Ray (CXR). First off, it's expected that ResNet-34 is used for feature extraction layer in lung disorders. Subsequently, the dataset is trained and tested using deeper network layers and improved feature extraction layers. ResNet-34 performs well in image classification, ImageNet features have been extracted from images. TRNN weights can be iteratively changed in accordance with training CXR images with a lower bias rate. EOL is a one-step optimization approach. National Institutes of Health (NIH) CXR dataset is collected from Kaggle and subjected to the EOL-TRNN model. With CXR images, the EOL-TRNN classifier may predict lung disease with a higher degree of accuracy than current techniques. Measures like precision, recall, Fβ-score, and accuracy are used to evaluate the performance of the approaches.