Prediction of Stroke Disease using Convolutional Neural Network and Multimodal Biosignals

Abstract

The high mortality and disability rates associated with strokes highlight the need of early diagnosis and preventative measures. Both ischemic and hemorrhagic forms of these illnesses need urgent care, although the two types differ in their specifics. In order to get expert medical help within the optimal treatment window, it is essential to recognize precursor symptoms as soon as possible. Nevertheless, prior research has mostly concerned itself with post-onset therapy rather than the identification of predictive signs. In this approach, a method for real-time stroke illness prediction using Convolutional Neural Network and multi-modal bio-signals are used The CNN-LSTM model demonstrates a satisfying accuracy of 99.15% utilizing raw data, and the system takes into account the convenience of senior patients to obtain high prediction accuracy. The suggested approach has the ability to identify and prevent strokes earlier than current imaging methods allow.