Deep CNN–Real ESRGAN: An Innovative Framework for Lung Disease Prediction

Abstract

The advent of Deep learning models led to an unprecedented change in the field of medical image analysis. Various models like Dense net, Mobile net, and Nas net achieve an accuracy of more than 80% but the lack of clean datasets makes it difficult for these algorithms to be effective in the classification of medical images. Data preprocessing techniques based on SRGAN (Super resolution generative adversarial network) have been used to improve the resolution of the image but they are still far from addressing general real-world degraded images. Our approach proposes to use Real ESRGAN (Enhanced super resolution generative adversarial network) on a lung disease dataset that includes six different types of lung disease. When preprocessed images with actual ESRGAN (Enhanced super resolution generative adversarial network) are used in deep CNN (Convolutional neural network) models, their classification accuracy improves. In our paper, three CNN models Mobile net, Nas net, and Dense net are combined with Real ESRGAN and their classification accuracy enhances above 90% and images which were not correctly classified with base models,are now classified with near to one probability.