A Speculative Approach For Brain Tumor Detection Using Image Processing

Abstract

Abstract— The brain is the essential organ within the human body, controlling and regulating all crucial existing functions, and a tumour is a mass of tissue formed through an accumulation of extraordinary cells that preserve to grow. A brain tumour is a tumour that has formed within the brain or has migrated there, for which no primary cause has been identified to date. Through the study it has been found that worldwide brain tumours make up only 1.8% of total reported tumours, the death rate of malignant brain tumours is very high because of their formation in the most essential organ of the body. Hence, there is a need to detect brain tumours at early stages with greater accuracy.
       We have thus proposed a system which will assess brain tumors from MRI scans for the management of brain tumor diagnosis. In this study, we will implement a model that, preprocesses the image, performs the post- processing by applying Morphological operation, segments the image using contour based edge detection and then finally classifies the tumors using EfficientNet which is a model of convolutional neural network architecture and gives the high degree of accuracy.