Deep Learning-based Precision Diagnosis of Lung Diseases on the Internet of Medical Things (IoMT)

Abstract

Lung disease is one of the common and severe pathological conditions that affect the respiratory system, causing respiratory illness and potential mortality. In recent times, deep learning paradigm based on the Internet of Medical Things (IoMT) platform has been adopted as a viable solution to address the challenges encountered in detection of lung diseases which are characterized by their diverse nature and the complexities associated with their diagnosis. In this work, we have proposed an approach that aims to achieve accurate prediction and analysis of lung diseases. The proposed research methodology presents a Deep Learning-based Accurate Lung Disease Prediction (DL-ALDP) model based on deep learning algorithms to enhance its predictive capabilities. The DL-ALDP framework integrates several preprocessing techniques, including Wiener filtering, optimized region growing method (ORGM)-based feature extraction, and Contrast limited AHE (CLAHE)-based segmentation. The accurate prediction of lung diseases is achieved by utilizing a Deep Neural Network (DNN) for classification purposes. The DL-ALDP technique, as suggested, attained a precision of 86.77%, sensitivity of 82.47%, specificity of 92.87%, accuracy of 92.08%, and F1 score of 89.42%. The findings of this research underscore the prospective utility of deep learning techniques in forecasting and analyzing lung ailments within the context of the IoMT platform. Through IoMT capabilities, healthcare practitioners can avail themselves of enhanced prognostic accuracy and timeliness, resulting in superior patient care and outcomes.