Abstract:

Because stroke illness often results in death or major disability, aggressive primary prevention and early diagnosis of prognostic signs are critical. Stroke illnesses are classified as ischemic or hemorrhagic, and they should be treated as soon as possible with thrombolytic or coagulant therapy. First, it is critical to notice the precursor symptoms of stroke in real time, which vary by person, and to give professional treatment by a medical institution within the appropriate treatment window. Prior research, however, has concentrated on creating acute therapy or clinical treatment recommendations after the onset of stroke rather than identifying predictive indicators of stroke. Image analysis, such as magnetic resonance imaging (MRI) or computed tomography (CT), has been utilised extensively in recent research to identify and predict prognostic signs in stroke patients. These approaches are not only difficult to identify early in real-time, but they also have drawbacks in terms of extended test times and expensive testing costs. In this research, we present a machine learning-based method for predicting and semantically interpreting stroke prognostic symptoms in the elderly utilising multi-modal bio-signals of electrocardiogram (ECG) and photoplethysmography (PPG) recorded in realtime. We devised and deployed a stroke disease prediction system with an ensemble structure that integrates CNN and LSTM to predict stroke illness in real-time while walking. The suggested system takes into account the ease of wearing bio-signal sensors for the elderly, and biosignals were captured while walking at a sample rate of 1,000Hz per second from the three electrodes of the ECG and the index finger for PPG. Real-time prediction of elderly stroke patients demonstrated good prediction accuracy and performance.

Description:

Deep learning, machine learning, electrocardiogram (ECG), photo plethysmography (PPG), multi-modal bio-signal, real-time stroke prediction, stroke disease analysis.