A Feature Fusion Based System for Brain Tumor Classification

Abstract

classifying brain tumors is an exclusive and difficult task in the field of clinical image analysis. Radiologists were able to reliably detect tumors using machine learning algorithms without the need for extensive surgery. However, a number of difficulties arise, including the difficulty in locating a specialist with competence in the field of identifying brain malignancies using images using deep learning models, and the main issue in the erection of the most effective deep learning system for diagnosing tumor cells. We used deep learning and adaptive algorithms to build a sophisticated and incredibly accurate system that uses feature fusion to automatically categorize brain tumors. The proposed framework extracts deep features from CNN architectures with varying depths and designs. The features from highest performing CNN architectures are then fused to form a single vector which is classified using SVM and KNN. The novel vector obtained the highest accuracy of 92% via the feature fusion method. Therefore, the suggested framework can be successfully used in clinical settings to categorize three different forms of tumors, namely gliomas, meningiomas, and pituitary tumors, from medical imaging.