This study aims to improve the jumping height for spiking in volleyball through skeleton-based action analysis using the Azure Kinect DK. Building on previous studies, this work focuses on the stride length, arm swing height, and spine angle during the jumping motion. Experiments are conducted to compare the efects of relatively better and relatively worse executions of these movements on jumping height. The results indicate that a relatively better stride length resulted in a significant increase of 187 mm in jumping height, while a relatively better arm swing height led to a 34 mm increase. However, variations in spine angle do not produce a significant diference in jumping height. These results suggest that optimizing stride length and arm swing height can efectively enhance the jumping performance of volleyball players.
for club activities [
In recent years, extensive research has been conducted on AI-based motion analysis in sports,
showcasing the potential of AI coaching [
This study aims to improve the jumping height for spiking in volleyball through skeletonbased action analysis using the Azure Kinect DK. By focusing on key factors such as stride length, arm swing height, and spine angle, this research seeks to determine how variations in these movements afect jumping height. Through this analysis, the study aims to provide coaches and players with more precise and actionable feedback, ultimately enhancing the efectiveness of volleyball training and performance. The rest of this paper is organized as follows: Section 2 introduces the method and algorithm in this study. Section 3 shows the experimental results. Section 4 gives a brief discussion and finally section 5 concludes this paper.
analyzed. The derivation of stride length was calculated as follows.
= √︀(1 − 2)2 + (1 − 2)2 + (1 − 2)2 The hand height during the backswing is derived from the equation of the plane of the floor and the distance between the joint points. The equation of the plane is obtained by the following equation.
+ + + = 0 The coeficients , , , and in this equation are expressed by the following equations (3)– (6) when there exist four 3-dimensional coordinates (1, 1, 1), (2, 2, 2), (3, 3, 3), (4, 4, 4).
= 2(3 − 4) + 3(4 − 2) + 4(2 − 3) = 2(3 − 4) + 3(4 − 2) + 4(2 − 3) = 2(3 − 4) + 3(4 − 2) + 4(2 − 3) = − 2(34 − 43) − 3(42 − 24) − 4(23 − 32) The distance D between a point and a plane is obtained from the equation of the plane shown in equation (2) and the 3-dimensional coordinates (1, 1, 1) as follows.
= |1 + 1 + 1 + | √2 + 2 + 2 The angle of the spine is obtained by the angle between three points. First, vectors ⃗, ⃗ are obtained from the coordinates of the three points. The vectors are obtained as follows. From these equations, the angle between the three points is obtained as follows. ⃗ = (1 − 2, 1 − 2, 1 − 2) ⃗ = (3 − 2, 3 − 2, 3 − 2) cos = ⃗ · ⃗
⃗ |⃗||| (8) (9) (10)
were conscious of spinal extension was 170°, and the average of the angles when the participants were conscious of bending was 171°. Thus, the diference in the mean of the angles of the two spinal columns is 1°. The mean of the highest reaching point was 2408 mm in the case of the bent spine and 2456 mm in the case of the extended spine. Therefore, the mean diference of the highest point reached was 48 mm, and the highest point reached was higher when the participants were conscious of extending the spinal column.
The analysis of stride length revealed a substantial diference in the highest arrival point, indicating its paramount importance in spiking performance. Specifically, a diference of 187 mm was observed, underscoring the critical role of stride length. However, it is noteworthy that while the analysis solely focused on the distance between the heels of both feet, future investigations should consider the timing and speed of the stepping motion to provide a comprehensive understanding.
In examining the backswing, our analysis of joint coordinates successfully identified significant variations in backswing heights. This underscores the validity of the analysis method employed in our study, highlighting its potential for enhancing spiking techniques.
Contrastingly, the analysis of spinal column angle revealed no significant diference between the two angles, with only a 1° disparity noted. This suggests limitations in Azure Kinect’s ability to accurately capture joint coordinates for the range of body movements evaluated in our study. Consequently, improvements in jumping motion may not be achievable through instructional interventions based solely on joint coordinate analysis using this methodology.
Moving forward, future research endeavors should explore alternative methodologies or technologies to overcome these limitations and further elucidate the intricacies of spiking techniques in volleyball.
In this study, Azure Kinect DK was utilized to capture joint coordinates during jumping motions in volleyball spikes, with a focus on stride length, backswing height, and spine angle. While analysis of joint coordinates for stride length and backswing height revealed diferences between the two motions, examination of spinal column angle did not yield any significant disparities. Although it was anticipated that improvement in jumping motion could be achieved through analysis of joint coordinates for stride distance and backswing height, no significant diference was observed in spinal column angle. Consequently, it is inferred that this method may not efectively enhance jumping motion. Future research endeavors aim to expand sample size and explore alternative analysis methods.
This research is partly supported by a research grant provided by the Shikino High-Tech Co., Ltd.