فهرست نشریات علمی دسترسی آزاد ایران

Background and Objectives: Wrestling is a high-intensity competitive sport with a high prevalence of knee injuries, among which anterior cruciate ligament (ACL) tears and other knee pathologies are common and clinically impactful. The wrestlers involved in this study ranged in age from 18 to 35 years old, including both professional and amateur athletes. All participants had achieved remarkable accomplishments in various national and international competitions, such as champions of the National Wrestling Championship and medalists of the Asian Wrestling Championship. These athletes sustained knee injuries during high-intensity training or competition, and were clinically diagnosed with ACL tears or other knee pathologies accordingly. Accurate diagnosis of ACL tears based on magnetic resonance imaging (MRI) is of critical importance for clinical decision-making and rehabilitation planning. However, the conventional manual analysis of MRI images is not only time-consuming but also prone to subjective deviations and diagnostic errors.    Methods: We propose a joint model combining an Improved Political Optimizer (IPO)—a metaheuristic optimization algorithm inspired by political competition and coalition strategies—with a Convolutional Neural Networks (CNN) to detect ACL tears in wrestlers. MRI images undergo standardization, normalization, and data augmentation before CNN automatically extracts deep features of the ACL region. IPO globally optimizes key CNN hyperparameters (e.g., learning rate, convolutional kernel size), enhancing both accuracy and efficiency.    Results: The IPO-optimized CNN achieved 93.5% accuracy, 91.8% sensitivity, and 94.7% specificity, while reducing training time by over 60% compared to traditional methods. Grad-CAM heatmaps confirmed that model attention aligns with expert-identified diagnostic regions.    Conclusion: The proposed IPO–CNN framework offers an accurate, efficient, and interpretable MRI-based solution for ACL tear detection, providing practical support for sports injury diagnosis and rehabilitation.