Detection of Brain Tumor Types based on Fanet Segmentation and GLRLM with Ensemble Learning

Enhancing the prognosis of brain tumors through early detection is crucial for enhancing patient survival through improved treatment results. It is a challenging task to manually assess the Magnetic Resonance Imaging (MRI) generated on a regular basis in the clinic. Thus, more accurate computer-aided techniques are desperately needed for early tumor diagnosis. Tumor identification, segmentation and classification procedures are part of computer-aided brain tumor diagnosis from MRI. In this model, uses the RS-ESIHE preprocessing approach to acquire enhanced MRI images. Pre-processed MRI is segmented using the FANET and segmented images are extracted using the GLRLM technique. The ensemble learning classifier is trained to predict different types of brain tumors. This proposed model achieves performance metrics of 98.5%, 97.5%, 96.8%, 1.5%, and 99% for Accuracy, precision, F1-Score, error and specificity. The comparisons are conducted between the assessed values and existing approaches like Mask-RCNN, DCNN, and SENET. Thus, the detection of brain tumor types based on FANET segmentation and GLRLM with an ensemble learning classifier performs better prediction than the existing model.