Investigation of hip flexibility training on dancesport optimization using machine learning video analysis

Ariticle ID: 348

Abstract

Dancesport, particularly the Paso Doble, requires high agility, coordination, and flexibility, especially in the hips. This study investigates the impact of an eight-week targeted Hip Flexibility Training (HFT) program on the performance of professional Paso Doble dancers. The need for this research stems from the lack of objective, data-driven evaluations in the field, where traditional methods rely heavily on subjective assessments. Previous studies have examined general flexibility in dance, but few have focused on the direct Biomechanical Effects (BF) and Physiological Effects (PE) of specific HFT on dancers. Further, such studies could not accurately measure hip joint movements and their coordination in order to achieve dance performance efficiency. The proposed study used motion-capturing devices to collect key movement data that impacts performance efficiency. The collected data is analyzed using the hybrid receptive field block (RFB) and residual network (ResNET) ML models to study the pre- and post-HFT results. Twelve highly trained dancers were assigned to have biomechanical and physiological metrics measured after and before the training. The data analysis has shown that there has been a significant increase in hip flexion from 65.4 ± 4.5° to 75.2 ± 3.7° (P < 0.05), hip extension from 25.3 ± 2.4° to 30.1 ± 2.1° (P < 0.05), and joint velocity from 1.18 ± 0.15 m/s to 1.32 ± 0.12 m/s (P < 0.05). Physiological metrics also showed improvements, such as a reduction in Oxygen Consumption (OC) from 2.02 ± 0.21 L/min to 1.85 ± 0.18 L/min (P < 0.05) and Energy Cost (EC) from 50.1 ± 7.2 kJ/min to 45.6 ± 6.4 kJ/min (P < 0.05).