Markerless analysis of front crawl swimming

Area of Science:

• Biomechanics
• Sports Science
• Motion Analysis

Background:

• Traditional swimmer motion analysis relies on manual video digitization, which is time-consuming and has low repeatability.
• There is a need for more efficient and reliable methods for quantitative analysis of swimming technique.

Purpose of the Study:

• To develop and evaluate a video-based, markerless system for analyzing arm movements in front crawl swimming.
• To assess the accuracy and reliability of the developed system compared to traditional manual digitization.

Main Methods:

• Modified a previously proposed method (Corazza et al., 2010) for underwater motion analysis.
• Used a markerless system to determine 3D coordinates of shoulder, elbow, and wrist joints for 5 sprint swimmers.
• Calculated wrist joint velocity and compared trajectory data with manual digitization (SIMI Reality Motion Systems GmbH).

Main Results:

• The system demonstrated good accuracy for wrist joint analysis, with Root Mean Square Distance (RMSD) values below 56mm.
• Reliability was found to be comparable to the inter-operator variability of manual digitization.
• The markerless system provides a promising alternative for quantitative, large-scale swimming motion studies.

Conclusions:

• The developed video-based, markerless system offers a more accurate and reliable method for analyzing swimmer motion.
• This technique addresses the limitations of traditional manual digitization, saving time and improving consistency.
• The system is well-suited for future quantitative research and large-scale studies in swimming biomechanics.