Deep-Learning Based Hybrid Model For The Classification of Lung Diseases

Lung diseases are one of the deadliest diseases which affects many people worldwide. These diseases have become more prevalent due to increase in air pollution. In this paper, we use the Shenzhen and Montgomery small volume datasets to solve the issue of classifying tuberculosis from chest X-ray pictures. The model that is proposed in this paper provides a classification model which employs two convolutional neural networks (Inception-V3 and Xception) that have been tested for lung disease classification using the ImageNet dataset. Furthermore, Squeeze and Excitation module are included in the suggested model for classifying X-ray of the chest into the following two categories: normal and Tuberculosis. To verify the effectiveness and efficiency of our suggested classification model, numerous experiments are conducted on the Shenzhen and Montgomery dataset. Further, the proposed model an accuracy of about 98% and 95% on Shenzhen and Montgomery datasets.