Machine Learning Approaches for Early Brain Stroke Detection Using CNN

Abstract

The rapid advancement in medical imaging technologies and the integration of artificial intelligence (AI) have
revolutionized diagnostic processes in healthcare. This study presents a novel tactic for the credentials of brain strokes
using CNN, leveraging the power of DL to boost indicative accurateness besides efficiency. Brain strokes are foremost
root of morbidity and mortality worldwide, necessitating timely and accurate diagnosis to improve patient outcomes.
The anticipated scheme mechanizes progression by employing a CNN-based model and haphazard woodland model to
analyze brain imaging data, specifically focusing on early detection and classification between stroke and non-stroke
conditions. The system was developed using a Flask web application framework, integrating a pre-trained CNN and
Random Forest model to process and classify brain images. The application facilitates user authentication, allowing
secure access to the predictive model. Users can upload brain metaphors, which stay formerly pre-processed and
analyze by the CNN model. To enhance the practical utility of scheme, information on possible causes and suggested
treatments for detected strokes is provided. Initial testing of the system demonstrated promising accuracy in stroke
exposure, prominence the probable of CNNs in medical diagnostics. The accurateness of the random forest model is
86% and the accurateness of the CNN model is 91.6%.As the result outcomes CNN model gives the better accuracy.