Özet:

Abstract Brain tumours are disreputably dangerous and difficult to treat. Brain tumours are identified by a laborious and error-prone process of manual visual inspection of pictures and manual marking of the suspicious areas by medical specialists. Magnetic Resonance (MR) images have had their ambiguity resolved in a more straightforward fashion. The work analyses the MRI images considered from public dataset providers. In recent years, researchers have proposed automating ways to detect brain cancers at an early stage. The tumour is a common malignant development with atypical features. Tumors of the brain are a form of abnormal growth of tissue in which cells multiply rapidly and out of control. Nature, origin, development rate, and maturity level are used to define its many varieties. Traditional methods of tumour detection are laborious, limited in their ability to effectively process vast amounts of data, and inaccurate. Hence, computer-aided diagnosis relies heavily on MRI's ability to automatically detect brain cancers. Variations in tumour location, shape, and size present a significant obstacle for brain tumour detection. The importance of early detection of brain cancers cannot be overstated. Methods based on computational intelligence can aid in the diagnosis and categorization of brain tumours with accurate grade detection that helps in proper treatment. To aid doctors in the early detection of malignancies with accurate grade detection, a Multi Level Weighted Group Feature Set based Dissimilar Region Detection (WGFS-DRD) for accurate grade detection in brain tumor detection. The proposed model is compared with the traditional models and the results represents that the proposed model performance is high.

Anahtar Kelimeler: