Methods for detecting brain tumours have been the primary emphasis of this work. Among the many vital vision applications in medicine, brain tumour detection stands out. An effort to unite bottom-up affinity-based segmentation methods with top-down generative model based approaches is initially introduced in this work via a survey of several well-known strategies for automated segmentation of diverse picture data. Investigating several methods for the effective detection of brain tumours is the primary goal of the effort. Poor quality photos, such as those with noise or low brightness, have been mostly disregarded by most current approaches. Additionally, objectbased segmentation has been largely disregarded in the majority of the current tumour identification research. Therefore, this study proposes a novel method to address the shortcomings of previous efforts. Results have been far better than those of a neural networkbased tumour detection method using this method. Wi