Published Paper Details:

The lungs are organs in the human respiratory system that function as an exchange of oxygen with carbon dioxide in the blood. The decline in air quality and COVID-19 in the world, including in India, has an impact on increasing the risk of lung disease. Lung disease that is often encountered is wet lung. Wet lung is a form of acute respiratory infection namely Lungs Diseases that attacks the lungs. The lungs are made up of tiny sacs called alveoli, which fill with water when a healthy person breathes. When a person has Lungs Diseases, the alveoli fill with pus and fluid, which makes breathing painful and limits oxygen intake. Early diagnosis of Lungs Diseases has a major impact on the life of a patient. Diagnosis of Lungs Diseases is generally done clinically (physical symptoms by a doctor). In addition, Lungs Diseases can also be diagnosed through chest radiographs, CT scans, and MRIs. Chest radiograph examination is one of the most frequently used medical imaging examinations because it is more affordable. Reading chest radiographs has drawbacks, namely it is difficult to detect disease, so it takes a long time before medical personnel or doctors diagnose the disease suffered by the patient. One method to overcome this problem is to classify chest radiograph images into certain classes using machine learning. The method used to process the data is multilayer perceptron (MLP). MLP has disadvantages for several types of data, especially for images, but MLP is not well adapted so that it loses spatial information contained in images. Therefore using multiple Convolutional Neural Network Architectures (AlexNet and ResNet) within which the importance of being able to identify the better architecture to obtain the results (diagnose Lungs Diseases) has been essential for the research performed, according to the antecedents of the authors proposed in the methodology, the best architecture analyzed is that of ResNet for having presented high performance in past research. Consequently, The joint classification (Resnet and Alexnet) has proven to be an innovative idea at the time of show a result of a predictive model, as it uses several classifications in turn to determine an end result. Which helped ensure the veracity of the output, which is very important to be able to determine if there is pneumonia in early stage, being this difficult to detect. Designing a Predictive Model for Early Detection of Lungs Diseases using Deep Learning and allowed this research to contribute providing a tool that serves as support for the diagnosis of this disease and reduce the high death rate that exists nationally and internationally.