A new systematic model for analysis and a hybrid fuzzy multimodality model for lung tumor prediction

Abstract

Lung cancer is one of the most commonly occurring diseases that ranked in the top of the present survey. Advancements in the medical field enable non-invasive methods of computerised diagnosis procedures and detection processes. Deep learning methods are already in evaluation by keeping the deep analysis on improving segmentation accuracy and prediction accuracy etc. The classification of tumour type depends on the quality of segmentation work and feature mappings. In this paper, we developed a robust model that classifies the types of tumours with improved accuracy but is also capable of detecting the early stages of cancer by detecting the unique hidden points of the image intensity in the lung images, etc. The system is comprised of a novel relative convergence technique for feature extraction technique to extract the infected area and its characteristic pixels to evaluate a unique feature mapping vector. The MSB feature mapping vectors are analysed with Hybrid Regress Fuzzy Net. The final result on whether a tumour is present in the CT image or normal depends on the three individual decisions made by the three algorithms mentioned. The accuracy of each algorithm is also considered for the probable decision-making. The performance measure of the entire proposed Hybrid Regress Net is evaluated through Accuracy, Precision, Recall and F1Score etc.