TeleStroke: real-time stroke detection with federated learning and YOLOv8 on edge devices

Journal of Real-Time Image Processing (2024) 21:121 https://doi.org/10.1007/s11554-024-01500-1 RESEARCH TeleStroke: real‑time stroke detection with federated learning and YOLOv8 on edge devices Abdussalam Elhanashi1 · Pierpaolo Dini1 · Sergio Saponara1 · Qinghe Zheng2 Received: 25 March 2024 / Accepted: 15 June 2024 © The Author(s) 2024 Abstract Stroke, a life-threatening medical condition, necessitates immediate intervention for optimal outcomes. Timely diagnosis and treatment play a crucial role in reducing mortality and minimizing long-term disabilities associated with strokes. This study presents a novel approach to meet these critical needs by proposing a real-time stroke detection system based on deep learning (DL) with utilization of federated learning (FL) to enhance accuracy and privacy preservation. The primary objective of this research is to develop an efficient and accurate model capable of discerning between stroke and non-stroke cases in real-time, facilitating healthcare professionals in making well-informed decisions. Traditional stroke detection methods relying on manual interpretation of medical images are time-consuming and prone to human error. DL techniques have shown promise in automating this process, yet challenges persist due to the need for extensive and diverse datasets and privacy concerns. To address these challenges, our methodology involves utilization and assessing YOLOv8 models on comprehensive datasets comprising both stroke and non-stroke based on the facial paralysis of the individuals from the images. This training process empowers the model to grasp intricate patterns and features associated with strokes, thereby enhancing its diagnostic accuracy. In addition, federated learning, a decentralized training approach, is employed to bolster privacy while preserving model performance. This approach enables the model to learn from data distributed across various clients without compromising sensitive patient information. The proposed methodology has been implemented on NVIDIA platforms, utilizing their advanced GPU capabilities to enable real-time processing and analysis. This optimized model has the potential to revolutionize stroke diagnosis and patient care, promising to save lives and elevate the quality of healthcare services in the neurology field. Keywords Stroke detection · Federated learning · Deep learning · Real time · Healthcare · NVIDIA platforms 1 Introduction Stroke is a leading cause of death and disability worldwide, making early detection and diagnosis crucial for improving patient outcomes. Rapid and accurate identification of stroke symptoms, along with timely diagnostic testing, are essential for initiating appropriate treatment and minimizing long-term consequences. This introduction will provide an overview of stroke detection and diagnosis, including the signs and symptoms of stroke, diagnostic tools and * Abdussalam Elhanashi 1 Ingegneria Informazione, University of Pisa, Pisa, Italy 2 School of Intelligent Engineering, Shandong Management University, Jinan 250357, Shandong, China techniques, and the importance of early intervention [1–4]. Stroke, also known as a cerebrovascular accident, occurs when the blood supply to the brain is interrupted, leading to damage or death of brain cells. There are two main types of strokes: ischemic stroke, caused by a blockage in a blood vessel supplying the brain, and hemorrhagic stroke, caused by bleeding into the brain. The common signs and symptoms of stroke can be remembered using the acronym FAST: face drooping, arm weakness, speech difficulty, time to call emergency services. Other symptoms include sudden numbness or weakness in the face, arm, or leg, especially on one side of the body; sudden confusion, trouble speaking or understanding speech; sudden trouble seeing in one or both eyes; sudden trouble walking, dizziness, loss of balance or coordination; and sudden severe headache with no known cause [5]. When a patient presents with symptoms suggestive of a stroke, healthcare providers must act quickly to confirm the Vol.:(0123456789) 121 Page 2 of 16 diagnosis and determine the type of stroke in order to initiate appropriate treatment. Several diagnostic tools and techniques are used in the evaluation of stroke patients. These include imaging studies such as computed tomography (CT) scans, magnetic resonance imaging (MRI), and angiography to visualize the brain to assess electrical activity in the brain; and blood tests to evaluate for potential causes of stroke such as high cholesterol, clotting disorders, or infection [6]. Early intervention is critical in the management of stroke as it can help minimize brain damage and improve patient outcomes. The “time is brain” concept emphasizes the importance of rapid assessment and treatment to preserve brain function. For ischemic strokes, timely administration of thrombolytic therapy (such as tissue plasminogen activator) or endovascular clot retrieval can help restore blood flow to the affected area of the brain. In cases of hemorrhagic stroke, prompt neurosurgical intervention would be necessary to control bleeding and reduce pressure on the brain. Despite advancements in stroke diagnosis, current diagnostic tools, such as CT and MRI, often fail to detect minor strokes and differentiate between ischemic and hemorrhagic strokes in the acute phase. There is also a lack of portable, rapid, and cost-effective diagnostic devices for use in pre-hospital settings, where early detection is crucial. Existing biomarkers lack specificity and sensitivity, limiting their clinical utility. The role of telemedicine in stroke diagnosis is underexplored, especially in remote areas with limited access to advanced medical facilities. It is necessary to bridge these gaps and improve diagnostic accuracy and timeliness, ultimately enhancing patient outcomes and reducing the burden on healthcare systems. Therefore, accurate and timely diagnosis is essential for guiding appropriate interventions and improving patient prognosis [7, 8]. In recent years, the rapid advancement of artificial intelligence (AI) and deep learning technologies has revolutionized various industries, including healthcare. These cutting-edge technologies have shown great promise in transforming the way medical diagnostics are conducted and in enhancing the delivery of e-healthcare services. By leveraging AI and deep learning, healthcare professionals can harness the power of data-driven insights to improve diagnostic accuracy, optimize treatment plans, and provide personalized care to patients. This introduction will explore the utilization of AI and deep learning in diagnostics and e-healthcare, highlighting their potential benefits and implications [9]. The integration of AI in healthcare has significantly impacted the way medical professionals approach diagnosis and treatment. AI algorithms have demonstrated the ability to analyze complex medical data, suc (...truncated)