An Effective Early Stage Detection of Lung Cancer Using Fuzzy Local Information cMean and GoogLeNet

Cancer is one of the main causes of death worldwide, accounting for an incredible 5 million fatalities per year. In this article, innovative machine learning algorithms are used to detect lung cancer at an early stage. To extract features, computed tomographic scan images were used. In the initial stage of lung nodule, preprocessing is accomplished for data cleaning and resizing of dataset. In the second stage, a set of features was recovered from the preprocessed images using Fuzzy Local Information cMean (FLIcM). Aside from this, deep features were retrieved and merged together for improved performance using GoogLeNet. To detect small cell lung cancer (SCLC), scans with no tumours after categorization using Sup-port Vector Machine (SVM) were enhanced using Contrasted Limited Adaptive Histogram Equalization (CLAHE) to recognise small cell lung cancers. Other than simple nodules, which are noncancerous cells, the suggested model has shown to be the most effective at detecting SCLC; as a result, we were able to reach a classification performance of 91.5 %. The suggested model improves classification performance by 3 % when employing a diffused feature set for early stage detection of SCLC, compared without using CLAHE.