MRI Brain Tumor Classification for Multi Labeled Medical Images using Machine Learning Techniques

The topics covered in this paper include brain tumors, pituitary and gliomas, how to classify images from magnetic resonance imaging (MRI) scans, The majority of MRI datasets that are currently accessible have some brain images that are not labeled, and many of the datasets are moderate in size. Our suggestion is to utilize deep supervised learning in order to fully utilize the labeled data. A novel label-based supervised learning framework was developed and integrated with Deep CNN features. In order to create consistent classifications for the sliced pictures from the same 3D brain scan, a new 3D-2D consistent constraint is implemented. After that, utilizing both labeled data sets, a deep learning classifier is developed to classify various Pituitary and glioma kinds. Testing the recommended technique on 3 datasets—pituitary, glioma, and non-tumor— we discovered that it completed well, with check accuracies of 99.2% in schooling and 99.10% in validation of the Gaggle Dataset.