An Enhanced Architecture for Brain Tumor Segmentation with a Novel Hybrid NF-RCNN and Adam Optimizer

Magnetic resonance imaging (MRI) of the brain for the purpose of manual tumor diagnosis is laborious and timeconsuming. A diagnostic by an expert is also necessary in every case. The need for accurate diagnosis and classification of brain tumors led to the development of several computer- controlled procedures. In order to segregate brain tumors, this research suggests a new filter. The approach used a hybrid model that included NasNet and FRCNN for segmentation of the brain tumor image, and it was optimized using Adam optimizer. For image filtering, SWT enhanced Median filter was utilized. Outperforming state-of-the-art approaches presented in literature, the suggested strategy shows promising outcomes. The proposed model segments brain tumor image with 92.19% accuracy on tumor region and 99.79% accuracy on background region.