تطوير وتقييم نظام ذكاء اصطناعي هجين لتشخيص كسور اليد: دراسة تكاملية بين الدقة التقنية والقبول السريري في مستشفى يفرن العام

Abstract

This study transitions from laboratory-based algorithmic development to the domain of clinical and field validation. It aims to evaluate the operational efficiency and professional acceptance of a hybrid Artificial Intelligence (AI) system specifically designed for diagnosing hand fractures. The significance of this research lies in its integrative approach, bridging the gap between high computational performance and the 

practical requirements of high-pressure healthcare environments, with a particular focus on addressing diagnostic challenges arising from complex hand anatomy and human fatigue.
The methodology employs a hybrid framework that integrates K-Means Clustering for enhanced image segmentation and feature extraction with Convolutional Neural Networks (CNN) for precise classification. The system was rigorously evaluated using a local radiological dataset sourced from Yefren General Hospital in Libya. To ensure practical applicability, a field study was conducted involving a survey of medical and technical staff ($n=22$) via a structured measurement tool to assess system attributes.
Results demonstrated an exceptional alignment between technical accuracy and clinical satisfaction. User interface (UI) interactivity achieved a significant mean score of 4.90/5, while participants affirmed the system's capacity to mitigate human error risks with a mean score of 4.18/5. The study concludes that the success of clinical decision support systems (CDSS) extends beyond software precision to encompass their integration into real-world workflows. Consequently, this model serves as a fundamental pillar for advancing radiological services in hospitals facing specialized staffing shortages.