Research Cell: An International Journal of Engineering Sciences

Kumud Sachdeva ¹, Amandeep Kaur ¹, Rajan Sachdeva ¹

Affiliations
1 Department of Computer Science and Engineering, Chandigarh University, Gharuan, Punjab, 140413, IN

Abstract

The primary causes of brain lesions are abnormalities in brain cells. Tumor develops in the brain as a result of these aberrant lesions. MRI and CT scans are two examples of medical imaging (CT) are the two different techniques for brain imaging scanning. The interior regions of the brain are scanned using MR imaging in this research project. There are two types of aberrant lesions in the brain: benign and malignant. Benign lesions can be treated with radiation therapy, whereas malignant lesions require adequate surgery performed by a radiologist with expertise. Uncontrolled cell proliferation in any area of the body is referred to as a tumor. Tumor come in a variety of forms, has unique traits, and call for a range of therapies. Brain tumors are a current problem. Malignant or metastatic brain tumors and primary brain tumors are two different categories. The metastatic or malignant tumors start as a cancer somewhere else in the body before spreading into the brain region, but the primary tumors start in the brain and have a tendency to stay there. The automatic tumor detection and segmentation method has therefore. In order to accurately classify and identify the tumor sections, three brain tumor segmentation methods are proposed in this study. First, a powerful brain tumor segmentation method is used by merging Convolution Neural Networks (CNN) and Multi Kernel K Means Clustering and Network (MKKMC). The proposed CNN-MKKMC technique uses the CNN algorithm to classify MR images into normal, IV, and abnormal categories. The next step is to separate the brain tumor from the aberrant brain imaging using the MKKMC algorithm. The accuracy, sensitivity, and specificity of the proposed CNN-MKKMC algorithm are assessed visually as well as objectively in comparison to the currently used segmentation techniques. The experimental findings show that the suggested CNNMKKMC technique produces greater segmentation accuracy for brain tumours while requiring less time. Both clinical datasets and publically accessible open access datasets are used to test the meningioma brain tumour detection approach.

Keywords

Neural Network, Deep Learning Neural Network, Convolution Neural Network, Multi Kernel K Means Clustering and Network, Multi Kernel K Means Clustering and Network.

Full Text