Skill Movement Trajectory Recognition of Freestyle Skiing U-Shaped Field Based on Deep Learning and Multitarget...

Area of Science:

• Sports Science
• Artificial Intelligence
• Computer Vision

Background:

• Freestyle skiing U-shaped venue skills demand high athlete proficiency.
• Traditional training methods may lack scientific optimization.
• Advancements in deep learning offer opportunities for enhanced sports training.

Purpose of the Study:

• To develop a scientific training method for freestyle skiing U-shaped venue skills.
• To integrate multi-target tracking algorithms with deep learning for motion capture.
• To improve athlete skill recognition and performance through advanced technology.

Main Methods:

• Combined convolutional neural networks (CNNs) and multi-target tracking algorithms for human action recognition.
• Utilized Long Short-Term Memory (LSTM) modules for analyzing freestyle skiing skills.
• Designed and conducted multi-target tracking dataset experiments and model updating experiments.

Main Results:

• The developed deep learning-based training method demonstrated a 14.48% improvement in skill scores compared to traditional methods.
• Experimental analysis and optimization refined the training approach.
• Professional students reported high satisfaction with the new training method.

Conclusions:

• The proposed deep learning framework, incorporating multi-target tracking, is effective for freestyle skiing U-shaped venue skills training.
• This approach offers a more scientific and data-driven method for skill enhancement.
• The findings suggest a promising direction for optimizing athletic training through AI.