Medical Image Classification Using DL-based Feature Extraction in IoMT

Aim: Recent advances in Artificial Intelligence (AI) and the addition of Deep Learning (DL) have made it possible to analyse both real-time and historical data from the Internet of Things (IoT). Recently, IoT technology has been implemented in healthcare schemes as IoMT to aid in medical diagnoses. Medical image classification is useful for predicting and identifying serious diseases at an early stage, which is crucial in the diagnostic process. Background: When it comes to managing, treating, and preventing illness, medical photographs are an essential element of a patient’s health record. However, it is a difficult issue in computer-based diagnostics to classify images using efficient characteristics. Objective: The patent study aimed to develop a deep learning-based classification model for feature extraction. Methods: Levy flight optimization is employed to pick the weight for the classification model optimally. At the end of the day, the optimal weight led to a better classification result and a higher degree of precision when analyzing medical photos for disease. Results: We tested the proposed results in MATLAB and compared them with conventional methods of classification. The suggested model’s best results include 97.71% accuracy on a brain dataset and 97.2% accuracy on an Alzheimer’s disease dataset. Conclusion: The proposed algorithm’s high rate of convergence proves that it can successfully balance the exploration and exploitation phases by avoiding capturing in local optimization and classifying thresholds rapidly. In light of the need for improved accuracy, precision, and computational speed in clinical picture classification, a novel approach based on soft sets has been presented. © 2024 Bentham Science Publishers.