An Advanced and Ideal Method for Tumor Detection and Classification from MRI Image Using Gaussian Mixture and RCNN

Extensive research and methodological improvements over time attest to the necessity of brain tumour detection and analysis for any indication system. Accurate tumour detection is essential for this study; hence an efficient automated method must be implemented Numerous segmentation algorithms have been developed to classify brain tumors with greater precision. Medical image processing experts agree that brain picture segmentation presents unique difficulties. A novel automatic detection and classification system is proposed in this work. The suggested method is broken down into several stages, including preprocessing of MRI images, segmentation of those images, extraction of features, and classification of those features. The noise in the MRI picture is reduced during the preprocessing stage by applying an adaptive filter. The grey level co-occurrence matrix (GLCM) is used for feature extraction, while the improved Means clustering (IMC) technique is used for image segmentation. To categorize MRI scans, we used a deep learning algorithm scan after features were extracted from the scans them as benign lesions, gliomas, meningiomas, or pituitary tumours. In order to accomplish the classification task, recurrent convolutional neural networks (RCNN) were utilised. For a given input dataset containing photographs of the brain, the proposed technique yields superior classification results. The investigations made use of MRI pictures from Kaggle's data sets, which included 464 testing sets and 3264 training sets. The outcomes show that the proposed strategy outperforms its predecessors. The suggested RCNN method is then compared to BP, VG-Net, and RCNN, three of the most popular classification techniques currently in use. The suggested classifier successfully identified 96.21 percent of brain tumour tissues in magnetic resonance imaging scans.