Leveraging large language models for gastrointestinal injury detection in athletes: a medical image analysis approach.
Authors
Affiliations (4)
Affiliations (4)
- Physical Education and Big Health, Yibin University, Level 2 units, Yibin, 644000, SiChuan, China.
- Physical Education and Big Health, Yibin University, Level 2 units, Yibin, 644000, SiChuan, China. [email protected].
- College of Sports, Nanchang Institute of Science and Technology, Nanchang, 330000, Jiangxi, China.
- Department of Sports and Leisure, Dongshin University, Naju-si, Republic of Korea.
Abstract
Injury detection and rehabilitation monitoring are critical components of sports medicine, particularly for gastrointestinal injuries that can impact athletic performance and health. Traditional approaches to injury assessment rely on manual medical imaging analysis and subjective clinical evaluations, which are often time-consuming, labor-intensive, and prone to interobserver variability. Recent advancements in artificial intelligence (AI), particularly large language models (LLMs) and deep learning techniques, offer new opportunities for enhancing medical image interpretation, predictive modeling, and real-time injury assessment. This study presents a novel AI-driven framework that integrates multimodal sports medicine data-including biomechanical signals, medical imaging, physiological indicators, and athlete performance metrics-to enhance the detection and management of gastrointestinal injuries in athletes. Our approach consists of a Biomechanical-Aware Neural Network (BANNet) for injury prediction and an Adaptive Rehabilitation and Performance Optimization Strategy (ARPOS) for personalized recovery interventions. BANNet employs hierarchical deep learning architectures to extract spatial-temporal patterns from diverse data sources, improving early injury detection and classification accuracy. ARPOS utilizes reinforcement learning, real-time sensor feedback, and individualized training adaptations to dynamically adjust rehabilitation protocols, optimizing recovery outcomes and minimizing reinjury risks. Experimental results demonstrate that our proposed framework significantly outperforms conventional methods in injury classification accuracy, rehabilitation efficiency, and clinical decision support. These findings highlight the potential of AI-powered systems in sports medicine, providing a scalable, interpretable, and data-driven solution for gastrointestinal injury detection, athlete performance optimization, and long-term health management.