A fine-grained attention and geometric correspondence model for musculoskeletal risk classification in athletes using multimodal visual and skeletal features
Researchers have developed ViSK-GAT, a novel deep learning framework designed to assess musculoskeletal risk in athletes. This multimodal model integrates visual and skeletal data, employing a Fine-Grained Attention Module for intra-modal feature refinement and a Multimodal Geometric Correspondence Module for cross-modal alignment. The system achieved over 93% in key metrics and demonstrated a low RMSE of 0.1205 and MAE of 0.0156, outperforming existing state-of-the-art methods. AI
IMPACT This model could enable earlier detection of potential injuries in athletes, leading to personalized training adjustments and injury prevention strategies.