Lung Cancer Detection Using Explainable Artificial Intelligence in Medical Diagnosis

Abstract

AI (artificial intelligence), IoT (internet of things), and CC (cloud computing) have all lately gained popularity in the healthcare industry, allowing radiologists to make better decisions. This research proposed a novel technique in lung tumor detection based on cloud-IoT in segmenting and classification using deep learning techniques. Using a cloud-based IoT module, the lung tumour dataset was gathered from multiple healthcare datasets. This data has been segmented using optimized fuzzy C-means neural network (OFCMNN). The pattern segmented area is attained by an optimized version. Then segmented pattern of lung cancer has been classified using ensemble of kernel multilayer deep transfer convolutional learning (KM-DTCL). The presented technique's performance was assessed using a benchmark image lung tumour dataset as well as lung MRI images. When compared to current strategies in the literature, the new method outperformed them in terms of accuracy, recall, precision, AUC, TPR, and FPR. The suggested method outperforms all of the photos in the application dataset in a variety of ways.