Next-Gen Stroke Detection: Neuroimage Analysis With Machine Learning

Exact and provoke stroke recognition is urgent for compelling treatment and working on tolerant results. In any case, customary demonstrative procedures frequently face difficulties in giving high precision and proficiency. This study presents an AI based technique for diagnosing strokes utilizing cerebrum X-ray examines.The approach leverages deep learning models, specifically MobileNet and ResNet, to classify images into two categories: Normal and Stroke. MobileNet was chosen for its lightweight structure, making it suitable for real-time use in settings with limited resources, while ResNet was selected for its powerful feature extraction, enhancing diagnostic accuracy. The models were trained on a diverse set of brain MRI images, and preprocessing methods were applied to optimize the quality of the data and improve the models’ performance. The results showed that MobileNet achieved an accuracy of 92%, and ResNet reached 94%, both outperforming the baseline CNN model, which had an accuracy of 85%. ResNet also showed superior precision, recall, and F1-score, making it ideal for applications requiring high accuracy. Meanwhile, MobileNet's balance of efficiency and accuracy makes it a suitable option for environments with limited computational resources. This exploration features the capability of profound learning in stroke location, giving medical care experts a solid device for early conclusion and opportune therapy. Future work will involve testing these models on larger datasets and conducting clinical trials to validate their effectiveness in real-world settings.