Pulmonary disease is a kind of disease that affects the lungs and other parts of the respiratory system and is mainly caused by smoking, asbestos, secondhand smoke, and other forms of air pollutants. Several types of pulmonary diseases are emphysema, fibrosis, pneumothorax, asthma, lung cancer, chronic obstructive pulmonary disease (COPD), and so on. Pulmonary diseases are otherwise known as lung disorders and respiratory diseases. Pulmonary diseases are predicted by several methods using x‐ray images and CT scan images. Some of the recent works predict only a specific disease, and optimal prediction is not yet achieved. Hence, we proposed a novel method known as African vulture optimization (AVO) algorithm‐based weighted support vector machine approach (w‐SVM). The proposed method in this article predicts emphysema, fibrosis, pneumothorax, and normal kinds from the NIH chest x‐ray dataset. The X‐images are preprocessed after data acquisition to obtain the desired size and to remove undesirable noise. The preprocessed images are then sent into the SVM for feature extraction, and the AVO is used to improve the SVM so that a kernel function may be obtained. The proposed w‐SVM effectively predicts the emphysema, fibrosis, pneumothorax, and normal classes from the dataset. The experimental analyses are conducted and compared with existing works and concluded that the proposed work outperforms other approaches in terms of accuracy, sensitivity, specificity, and Matthews's correlation coefficient, prediction time, and modeling time.

No abstract

Nowadays, deep learning plays a vital role behind many of the emerging technologies. Few applications of deep learning include speech recognition, virtual assistant, healthcare, entertainment, and so on. In healthcare applications, deep learning can be used to predict diseases effectively. It is a type of computer model that learns in conducting classification tasks directly from text, sound, or images. It also provides better accuracy and sometimes outdoes human performance. We presented a novel approach that makes use of the deep learning method in our proposed work. The prediction of pulmonary disease can be performed with the aid of convolutional neural network (CNN) incorporated with auction-based optimization algorithm (ABOA) and DSC process. The traditional CNN ignores the dominant features from the X-ray images while performing the feature extraction process. This can be effectively circumvented by the adoption of ABOA, and the DSC is used to classify the pulmonary disease types such as fibrosis, pneumonia, cardiomegaly, and normal from the X-ray images. We have taken two datasets, namely the NIH Chest X-ray dataset and ChestX-ray8. The performances of the proposed approach are compared with deep learning-based state-of-art works such as BPD, DL, CSS-DL, and Grad-CAM. From the performance analyses, it is confirmed that the proposed approach effectively extracts the features from the X-ray images, and thus, the prediction of pulmonary diseases is more accurate than the state-of-art approaches.