Health Literacy In Kanyakumari District And Factors Influenzing The Effectiveness Of Health Literacy Over Physical Activity
Main Article Content
Abstract
Past decades show the failures of lung cancer detection. Lung cancer is the most agonizing disease for humans and many of them do not get cured until the proper treatment [13]. Deep learning Techniques provide an efficient way for a radiologist to analyse the lung images properly and open the right path for accurate Segmentation and Classification. This paper proposed automatic segmentation by improved thresholding technique and evaluated by Dice similarity coefficient, Structure Similarity Index (SSIM), and Feature similarity Index (FSIM). Feature extracted by GLCM for classification and Classified by Neuro-fuzzy classification and RESNET 50 and compared these results with previous research by evaluation metrics Accuracy, Area under the ROC Curve, Sensitivity, Precision, and F1 Score and recommend better classification techniques for lung CT images. This model experiments with the images of the LIDC-IDRI dataset and achieves a 98.99 % Dice similarity coefficient, SSIM as 99%, FSIM as 97.11% for segmentation and classification achieves Accuracy as 99.78%, Area under the ROC Curve as 99.27%, Sensitivity as 99.13%, Precision as 99.01%, and F1 Score as 99.76%. This model suggests a better preprocessing technique denoising autoencoder for medical images and improved dice similarity coefficient for automatic segmentation of lung CT images and also reduce the model loss of RESNET50 in classification of lung CT image classification and attain improved result than [7].