Validation of an Automated Step Length Measurement Method in Sprinting Athletes Using Computer Vision and Pose Estimation

Sprinting is a high-intensity anaerobic exercise requiring power, speed, and efficiency to improve performance. Understanding the biomechanics of running, specifically step length, is crucial in this pursuit. Traditionally, step length has been measured manually by measuring the distance between successive foot strikes. However, with advancements in technology, computer vision techniques and pose estimation using Media Pipe's model can automatically calculate step length in real-time. This research article compares manual and automatic step length measurement accuracy and efficiency using Media Pipe for sprinting. To evaluate the proposed approach, data from 47 subjects running at their preferred pace on a 400-meter outdoor track were analysed and compared to results obtained through Kinovea software, a commonly used motion analysis software. The study findings demonstrate that the proposed method for step length determination is valid and precise. The Pearson correlation coefficient revealed a strong positive correlation of 0.921, and the Intra-class correlation coefficient (ICC) agreement of average measures is 0.959, indicating a high level of consistency between measurements. The Standard Error of Measurement (SEM) values of 68cm and Minimum Detectable Change (MDC) value of 1.89cm show that the proposed method offers high precision. In conclusion, this study highlights the effectiveness of using Media Pipe for implementing a computer vision-based approach to accurately and efficiently measure step length during sprinting.