A new improved deep learning method for brain tumor segmentation and classification.

The fascinating field of tumor segmentation and classification determines the tumor's level by differentiating between cancerous and non-cancerous cells. Due to the enormous datasets and varying image sizes, MRI segmentation presents a challenge. Numerous techniques were developed in the literature to categorize brain tumors; however, these techniques were unable to provide improved categorization due to their accuracy and poor decision-making. In order to improve classification accuracy and decision-making effectiveness, this work introduces an improved deep learning mechanism known as Mayfly optimization based Deep Convolution Neural Network (Mayfly-DCNN). Before going through the segmentation process, the input MRI pictures are pre-processed. The segmentation method is carried out using a fuzzy deformable fusion model using a Mayfly-based sine cosine algorithm. After that, statistical features like mean, variance, and skewness are used to extract features. The DCNN, which classifies brain tumors using the retrieved features, is trained using the Mayfly-based sine cosine optimization. The experimentation is carried out using MRI scans from the BRATS database and SimBRATS, and the proposed method has proven to perform better than the state-of-the-art methods in terms of accuracy.