Improved feature extraction and selection method for brain tumor detection

Abstract

Medical Image processing is a vital research area in the field of medical and computing technology. Among this Feature extraction and feature selection are the main tasks in pre-processing of info logs to recognize digital protection threats and attacks while using machine learning. With regards to the investigation of heterogeneous data got from various sources, these tasks are figured out to be time-consuming and hard to be overseen effectively. Feature selection and feature extraction techniques have dissected with the end goal of how adequately these techniques can be utilized to accomplish elite of learning algorithms that eventually improves prescient accuracy of classifier. In this phase, Feature Extraction is done by the Wavelet-Based Gray-Level Run-Length Matrix (GLRLM) and the extracted feature set is additionally refined through feature selection. After this Feature Extraction, naive bayes wrapper classifier is used for covering feature selection. Proposed strategy was utilized for feature selection are useful in distinguishing brain tumor where is actually found.