Feature Based Fusion Model for Automated Detection of Brain Diseases from MRI Images

Due to the increasing rate of disease afflictions among people lately, the need for automatic illness diagnosis systems is more imperative. Most of the automatic illness diagnosis systems aim to support the physician in disease screening and decision-making. Furthermore, an immense amount of research is done on the human brain, a complicated and important organ, using imaging methods like Magnetic Resonance Imaging (MRI) to look at brain activity and uncover abnormalities. When compared to other imaging modalities, MRI stands out for its superior soft tissue contrast and safety profile. This research aims to develop a computer-aided disease diagnosis system to classify brain MRI images with high accuracy. The research significantly contributes to the following: developing an ensemble model based on deep learning and proposing a new feature selection technique that employs a framework of rank-based correlation and entropy. The framework’s final step uses an ensemble learning process to classify the extracted features. The stages of the suggested framework are as follows: (i) gathering and resizing images; (ii) deep feature extraction using the pre-trained networks; (iii) handcrafted feature extraction; (iv) serial feature concatenation; (v) finest feature selection using entropy and rank-based correlation and (vi) classification using a voting classifier. The suggested approach intends to improve the effectiveness and accuracy of brain disease diagnosis, opening the door for early detection and prompt intervention by integrating MRI imaging modality and AI algorithms. Experimental investigations conducted using MATLAB software demonstrate promising results in the preprocessing of MRI pictures and the detection of brain diseases through the proposed fusion model by achieving an accuracy of 89.87%. The findings of this study emphasize the significance of AI-based approaches in automated brain disease detection, offering a valuable contribution to the field of medical imaging and diagnostics.