This article presents an innovative approach to segmenting brain tumours in human brain MRI scans. This paper proposes a methodology that combines the usability of diffusion filtering and a multi-thresholding Otsu technique in improving precision and reliability of tumour segmentation. Here, the results of this paper show how useful the approach may be, thus potentially an influencer in medical imaging diagnoses and treatment planning. The proposed methodology is compared with other existing methods for segmenting brain tumours from MRI and shows satisfactory results. We use these techniques to bend the active contour towards the tumour walls. The suggested strategy leverages the adaptability of active contours and the precision of a pre-fitting mask to enhance the accuracy and efficiency of the tumour border delineation process. The recommended method yields satisfactory results when compared to existing pre-fitting mask procedures, and the BraTS dataset validates its effectiveness.