Hiroki Ozaki
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Hiroki Ozaki has a total of 25 co-authors. Here are the top co-authors by number of shared publications:
- Asim Munawar
IBM Research AI IBM Research AI IBM Research AI IBM Research AI IBM Research Japan Institute of Sports Sciences Japan Institute of Sports Sciences Japan Institute of Sports Sciences Japan Institute of Sports Sciences Japan Institute of Sports Sciences
1 shared publication - Daiki Kimura
IBM Research
1 shared publication Erina Kurosaki
1 shared publicationHirotaka Nakashima
1 shared publication- Keita Watanabe
Faculty of Culture and Sport Policy, Toin University of Yokohama, Yokohama 225-8503, Japan
1 shared publication Kenny J. Z. Tan
1 shared publication
Latest Publications
- 1 January 2026
- Published by Human Kinetics in International Journal of Sports Physiology and Performance
- Vol. -1 (aop) , 1-7
Physiological Responses and Intensity Profiles During Simulated Match Play in World-Class Wheelchair Tennis Players
Purpose : This study aimed to clarify the physiological demands during simulated tennis matches on a hard court in world-class wheelchair tennis players. Methods : Two male and 4 female wheelchair tennis players completed 2 tests: an arm-cranking test in the laboratory and measurement of physiological responses during a wheelchair tennis match on a hard court. In the arm-cranking test, oxygen uptake ( ), heart rate (HR), and blood lactate concentration were determined to evaluate the maximum values of ( ) and HR (HR max ), their values at the lactate threshold (LT) ( and HR LT ), as well as the onset of blood lactate accumulation (OBLA) ( and HR OBLA ). On the court, and HR during matches were determined and expressed relative to the corresponding values determined in the laboratory ( and %HR max , respectively). We also calculated the cumulative times spent in 3 intensity zones, defined as zone 1 (≤LT), zone 2 (>LT to Results : The median and %HR max were 56.1% and 72.7%, respectively. was significantly positively correlated with zone 1 ( τ b = .733). The median was significantly higher in service than return games (service: 22.5 [19.2–23.4] mL/kg/min, return: 18.8 [18.0–21.3] mL/kg/min), which corresponded to significantly shorter and longer cumulative times spent in zones 1 and 2, respectively. Conclusions : World-class wheelchair tennis players played most of the match at moderate intensities between LT and OBLA, and a higher percentage of at LT was associated with longer time spent in lower intensity zones during matches.
- 1 January 2026
- Published by Human Kinetics in International Journal of Sports Physiology and Performance
- Vol. -1 (aop) , 1-7
Physiological Responses and Intensity Profiles During Simulated Match Play in World-Class Wheelchair Tennis Players
Purpose : This study aimed to clarify the physiological demands during simulated tennis matches on a hard court in world-class wheelchair tennis players. Methods : Two male and 4 female wheelchair tennis players completed 2 tests: an arm-cranking test in the laboratory and measurement of physiological responses during a wheelchair tennis match on a hard court. In the arm-cranking test, oxygen uptake ( ), heart rate (HR), and blood lactate concentration were determined to evaluate the maximum values of ( ) and HR (HR max ), their values at the lactate threshold (LT) ( and HR LT ), as well as the onset of blood lactate accumulation (OBLA) ( and HR OBLA ). On the court, and HR during matches were determined and expressed relative to the corresponding values determined in the laboratory ( and %HR max , respectively). We also calculated the cumulative times spent in 3 intensity zones, defined as zone 1 (≤LT), zone 2 (>LT to Results : The median and %HR max were 56.1% and 72.7%, respectively. was significantly positively correlated with zone 1 ( τ b = .733). The median was significantly higher in service than return games (service: 22.5 [19.2–23.4] mL/kg/min, return: 18.8 [18.0–21.3] mL/kg/min), which corresponded to significantly shorter and longer cumulative times spent in zones 1 and 2, respectively. Conclusions : World-class wheelchair tennis players played most of the match at moderate intensities between LT and OBLA, and a higher percentage of at LT was associated with longer time spent in lower intensity zones during matches.
- Open Access
- 7 May 2020
- Published by MDPI AG in Applied Sciences
- Vol. 10 (9) , 3256
Aerodynamic Characteristics of New Volleyball for the 2020 Tokyo Olympics
The pattern of a modern volleyball is greatly different from that of a conventional volleyball, with several changes being made to the shape and design of the surface on the ball. Furthermore, at the 2020 Tokyo Olympics, a new volleyball (V200W; Mikasa) with 18 panels will be shown as the official ball. Therefore, this study compared the basic aerodynamic characteristics of conventional volleyballs with those of new designs in a wind tunnel. We used three full-size FIVB (Fédération Internationale de Volley-Ball) official volleyballs (V5M5000; Molten, MVA200; Mikasa and V200W; Mikasa) to determine the aerodynamic forces acting on each ball. The results indicate that the critical Reynolds number (Recr) differed depending on the ball types and their orientations. The Recr for the Molten ball (conventional) was determined to be ~3.4 × 105 (Cd = 0.17) on panel orientation A and ~2.7 × 105 (Cd = 0.14) on panel orientation B. Moreover, the Recr for the conventional Mikasa ball was determined to be ~2.6 × 105 (Cd = 0.14) on panel orientation A and ~3.0 × 105 (Cd = 0.13) on panel orientation B. On the other hand, the critical Reynolds number for the new volleyball (V200W) was ~2.9 × 105 (Cd = 0.17) in the panel orientation A and ~2.6 × 105 (Cd = 0.15) in panel orientation B. From these results, it can be hypothesized that, during a float serve, the flight trajectory will change depending on the type of volleyball and their orientation.
- 7 May 2020
- Published by MDPI AG in Applied Sciences
- Vol. 10 (9) , 3256
Open AccessAerodynamic Characteristics of New Volleyball for the 2020 Tokyo Olympics
The pattern of a modern volleyball is greatly different from that of a conventional volleyball, with several changes being made to the shape and design of the surface on the ball. Furthermore, at the 2020 Tokyo Olympics, a new volleyball (V200W; Mikasa) with 18 panels will be shown as the official ball. Therefore, this study compared the basic aerodynamic characteristics of conventional volleyballs with those of new designs in a wind tunnel. We used three full-size FIVB (Fédération Internationale de Volley-Ball) official volleyballs (V5M5000; Molten, MVA200; Mikasa and V200W; Mikasa) to determine the aerodynamic forces acting on each ball. The results indicate that the critical Reynolds number (Recr) differed depending on the ball types and their orientations. The Recr for the Molten ball (conventional) was determined to be ~3.4 × 105 (Cd = 0.17) on panel orientation A and ~2.7 × 105 (Cd = 0.14) on panel orientation B. Moreover, the Recr for the conventional Mikasa ball was determined to be ~2.6 × 105 (Cd = 0.14) on panel orientation A and ~3.0 × 105 (Cd = 0.13) on panel orientation B. On the other hand, the critical Reynolds number for the new volleyball (V200W) was ~2.9 × 105 (Cd = 0.17) in the panel orientation A and ~2.6 × 105 (Cd = 0.15) in panel orientation B. From these results, it can be hypothesized that, during a float serve, the flight trajectory will change depending on the type of volleyball and their orientation.
- Open Access
Unsupervised Temporal Feature Aggregation for Event Detection in Unstructured Sports Videos
Image-based sports analytics enable automatic retrieval of key events in a game to speed up the analytics process for human experts. However, most existing methods focus on structured television broadcast video datasets with a straight and fixed camera having minimum variability in the capturing pose. In this paper, we study the case of event detection in sports videos for unstructured environments with arbitrary camera angles. The transition from structured to unstructured video analysis produces multiple challenges that we address in our paper. Specifically, we identify and solve two major problems: unsupervised identification of players in an unstructured setting and generalization of the trained models to pose variations due to arbitrary shooting angles. For the first problem, we propose a temporal feature aggregation algorithm using person re-identification features to obtain high player retrieval precision by boosting a weak heuristic scoring method. Additionally, we propose a data augmentation technique, based on multi-modal image translation model, to reduce bias in the appearance of training samples. Experimental evaluations show that our proposed method improves precision for player retrieval from 0.78 to 0.86 for obliquely angled videos. Additionally, we obtain an improvement in F1 score for rally detection in table tennis videos from 0.79 in case of global frame-level features to 0.89 using our proposed player-level features. Please see the supplementary video submission at https://ibm.biz/BdzeZA.
Open AccessUnsupervised Temporal Feature Aggregation for Event Detection in Unstructured Sports Videos
Image-based sports analytics enable automatic retrieval of key events in a game to speed up the analytics process for human experts. However, most existing methods focus on structured television broadcast video datasets with a straight and fixed camera having minimum variability in the capturing pose. In this paper, we study the case of event detection in sports videos for unstructured environments with arbitrary camera angles. The transition from structured to unstructured video analysis produces multiple challenges that we address in our paper. Specifically, we identify and solve two major problems: unsupervised identification of players in an unstructured setting and generalization of the trained models to pose variations due to arbitrary shooting angles. For the first problem, we propose a temporal feature aggregation algorithm using person re-identification features to obtain high player retrieval precision by boosting a weak heuristic scoring method. Additionally, we propose a data augmentation technique, based on multi-modal image translation model, to reduce bias in the appearance of training samples. Experimental evaluations show that our proposed method improves precision for player retrieval from 0.78 to 0.86 for obliquely angled videos. Additionally, we obtain an improvement in F1 score for rally detection in table tennis videos from 0.79 in case of global frame-level features to 0.89 using our proposed player-level features. Please see the supplementary video submission at https://ibm.biz/BdzeZA.