Early and accurate diagnosis of stroke in emergency medicine is critical to ensuring timely treatment and reducing long-term patient disability or mortality. However, the complex and time-sensitive nature of stroke symptoms, combined with the high-pressure environment of emergency departments, often leads to diagnostic delays or errors. This paper proposes an AI-powered Clinical Decision Support System (CDSS) designed to assist clinicians in the early detection of stroke using structured patient data such as vital signs, medical history, and neurological observations. The system leverages machine learning techniques to classify patient cases as likely stroke or non-stroke, providing real-time diagnostic support during triage or initial evaluation. To evaluate the effectiveness of the proposed approach, we develop a simulation environment that mimics emergency department workflows and incorporates both synthetic and publicly available clinical datasets. Performance is assessed using standard metrics including accuracy, sensitivity, specificity, and area under the ROC curve (AUC). The results demonstrate the potential of AI-driven decision support to enhance diagnostic accuracy for stroke patients in emergency settings.

An essential criterion in content-based image retrieval is its accuracy. High accuracy depends heavily on the quality of feature extraction and dimension reduction methods. To achieve this aim, this paper integrates Visual Geometry Group (VGG) 16-based feature extraction with Autoencoder-based dimensionality reduction, along with Grey Wolf Optimization (GWO), to enhance the K-means clustering algorithm. Initially, 4096 deep features were extracted from images using VGG16. These features were then reduced to 10, 15, and 30 latent features using an Autoencoder (AE) that captures complex data structures more effectively than linear approaches. Moreover, the clustering stage was further improved by applying GWO to promote the initialization of K-means centroids. The proposed method was assessed on two datasets: Corel-1k (10 classes) and UC Merced Land Use (UCM) (21 classes). Further, it was compared with the Principal Component Analysis (PCA)-based method and other recent deep learning approaches. The experimental results revealed that the proposed method outperformed other studies in both image retrieval accuracy by mean average precision, and clustering quality, assessed by Purity, Adjusted Rand Index (ARI), and Normalized Mutual Information (NMI). These results confirm that feature extraction using nonlinear dimension reduction by AE yields more discriminative features than PCA and other approaches in all metrics.

Background: Fuzzy Logic (FL) is increasingly used in medical domains as it provides a robust framework for dealing with the uncertainty and subjectivity often encountered in clinical environments. With the increasing complexity of medical data and the growing need for interpretable decision-making, fuzzy systems have been widely adopted across diverse healthcare domains. Objective: This paper contributes a broad and well-structured survey of fuzzy logic techniques applied in diverse healthcare scenarios. It categorises the literature based on application domains, methodologies, data sources, and implementation tools while identifying common challenges and proposing future research directions. Methodology: A Systematic Literature Review (SLR) was conducted using five major databases: IEEE Xplore, Scopus, ScienceDirect and Web of Science. Articles published between 2017 and 2025 were filtered through established eligibility criteria to ensure relevance.The selected studies were classified into six major healthcare categories: diagnosis, monitoring, treatment recommendation, risk prediction, supporting infrastructure, and review papers. Results: The review includes over 147 studies, revealing that diagnosis is the most frequently targeted application area, with fuzzy inference systems and hybrid models being the most commonly used techniques. While FL systems show strengths in interpretability and adaptability, many suffer from a lack of clinical validation, limited datasets, and scalability concerns in real-world deployment. Discussion: The findings suggest that although FL continues to offer significant value in modelling clinical uncertainty and facilitating explainable reasoning, the field remains fragmented. There is limited standardisation across implementations, and many studies lack real-time capability or integration with hospital systems. Furthermore, most models rely on expert-driven rule design rather than data-driven tuning, limiting their adaptability to evolving medical knowledge.