Brain Health Prediction using CNN-YOLOv11 Deep Learning Model

Abstract

When it comes to medical imaging and diagnostics, the capacity to interpret brain scans correctly and predict the neurological health conditions has become one of the key tasks that can potentially save lives. From early prediction of brain tumors to identifying degenerative disorders, automated Brain Health Prediction (BHP) system transforming the way healthcare professionals interpret complex medical data. Object Detection algorithms, such as YOLOv11 (You Only Look Once), have become popular due to their fastness and the capability to detect abnormalities in brain MRI. Hybrid models are under development by integrating the advantages of YOLOv11 to detect and CNN to classify to increase diagnostic accuracy and efficiency. CNN model is validating over various epochs between 1 and 20 that have an accuracy of 99.54% and the YOLOv11 model also demonstrated the accuracy of 0.93 that is efficient than other earlier models. The proposed model is a combination of Convolutional Neural Network (CNNs) to extract and classify features effectively, and the YOLOv11 to detect abnormal brain areas fast and accurately. The sample of healthy and diseased MRI brain images was selected and data augmentation was done. The CNN model achieved a high classification of 96.64% after 20 epochs and the YOLOv11 model achieved a validation of 93% and is faster and more accurate in comparison to other models. The results validate the proposed model potentially to enhanced clinical workflow, reduce diagnostic delay and support early detection of neurological disorders.