Prediction and Classification of Normal and Abnormal Lungs Images Using Deep Learning Model

ABSTRACT

The prediction and classification of images are still need more accuracy especially in the field of medical image classification required more and more to find out the exact classification by using image processing technique and apply the deep learning model for the better result in this sector. In this paper mainly work carried out in the classification of normal or abnormal of images if abnormal images as the classification further again apply the deep learning model to classify the category of the lungs images like Covid, Lungs opacity and viral pneumonia. The system has an input image as 200 X 200 and the entire image taken from kaggle open database for this entire research work. The system initially classify according to disease to train the model for the classification, here both the training and testing phase will be available, therefore 75% of training data and remaining 25% of testing will used in this process based on the existing system availability VGG16 and sequential model using keras were implemented in this model for the classification of normal and abnormal of disease affected human lungs images and the result justify that sequential model is an efficient model compare to existing system as well as implemented VGG16 in terms of time and accuracy are the factor taken into account for the justification and sequential model produce 98% of accuracy with less time consumption comparing to available data also precision, recall and F1-score also calculated to strongly recommend this model for the better classification of human lungs using the deep learning model with more accuracy with less time consumption. Copyright © 2022 Wolters Kluwer Medknow Publications. All rights reserved.