Sports movements need to be closely observed and modified to improve athletic performance and reduce the risk of injury. Conventional evaluation techniques are challenging to scale and rely on subjective opinion. Our solution to this problem is the Deep Learning-based Standardized Assessment and Correction System (SACS), which analyzes motion video sequences of athletes, detects biomechanical irregularities, and automatically suggests corrective actions using a CNN–LSTM architecture. To provide real-time, individualized training recommendations, the system integrates position estimation, biomechanical feedback, and multi-angle video inputs. SACS outperformed previous techniques in precision, recall, and F1-score, achieving 96% accuracy in recognizing and correcting movement patterns when tested on a large sports dataset. These findings show how SACS enables objective, real-time evaluation and feedback, helping coaches and athletes across a variety of sports enhance performance and reduce injury risk.
Tian et al. (Tue,) studied this question.