Improved Classification System of Brain Tumor from MRI Images Using Swarm Intelligence Based Convolution Neural Networks
Abstract
Since human brain tumours are the most common form of tumour affecting both men and women, image processing-based medical diagnosis has seen a significant rise in recent years. It is, however, curable if it is detected and classified at an initial stage. Researchers have been working on a complex mechanism for classifying human brain tumours for a long time. The use of image processing to classify tumours from Magnetic Resonance Imaging (MRI) is common in the detection of brain disorders. So, utilizing Convolution Neural Network (CNN) as a deep learning, a swarm Intellect -Based Improved Brain Tumor Segmentation and Classification (I-BTSC)model is built from MRI images in this research work. First, we use the Swarm-based Grasshopper Optimization (SGO) algorithm to segment tumour regions using the conventional K-means concept. Following tumour region segmentation, feature extraction is used to determine the feature collection, and the fitness role is again served by SGO-based feature selection. Finally, CNN is used as a deep learning methodology to train the suggested I-BTSC system, which is then calculated. The suggested I-BTSC system is simulated using a publicly accessible “Contrast-Enhanced MRI Dataset”, and the outcomes show that the suggested model is more efficient than previous research work interms of performance measures.