Computer Aided Diagnosis for Brain Tumor Segmentation using Fine Tuned Convolutional Neural Network

The human brain is the most vital and complicated organ in the body, restricting the entire organism. Correctly diagnosing a brain tumor is difficult due to its complexity. An uncontrolled cell proliferation development of brain cells is one of humanity's most lethal diseases in history. Several image processing techniques are applied for diagnosis, therapy, and surgery planning. MR brain imaging techniques play a critical role in diagnosing various brain ailments. The suggested methodology's primary purpose is to segment the MRI brain image's tumor portion effectively. In this research, Convolutional Neuronal Networks (CNN) is used to perform brain tumor segmentation. We developed a novel optimization technique based on the firefly algorithm to maximize segmentation efficacy. The proposed firefly algorithm optimizes the CNN model's essential hyper parameters. Many experiments are conducted to demonstrate the suggested research's tumor segmentation efficiency. The experimental findings suggest the superiority of the proposed method. According to the experimental results, our proposed segmentation system achieves a 98 percent accuracy, a 97 percent precession, and a 99 percent recall rate.