Proceedings of the 14th International Symposium on Computer Science in Sport (IACSS 2023)

Table tennis is a popular sport worldwide, especially in China. To win in a game, most amateurs, professional athletes, and coaches have to watch and analyze competition videos to accumulate experiences and lessons for improving skills. However, most frames in a table tennis competition video are useless for technical and tactical analysis, requiring the video to be segmented into pieces and to keep pieces with effective playing actions for analysis, which are called effective rounds. Nowadays, this kind of segmentation works are performed manually, which is expensive and inefficient. To solve this problem, we propose and implement an automatic effective round segmentation method for table tennis based on machine learning techniques. In this method, we design an action recognition module and a key frame discrimination module to identify specific key frames of the video stream. Based on these key frames, we design an automatic segmentation module to combine effective frames as effective rounds of a competition video. To validate the proposed method, we build a human action dataset, a table tennis key frames dataset, and a competition video dataset, and conduct extensive experiments. The evaluation results with high-performance indicators demonstrate the effectiveness and efficiency of the proposed method.

Spatial organization of players in invasion games, i.e., formation is an important tactical feature. In handball defensive formations mainly include 6:0, 5:1, and 3:2:1 with different numbers of players acting on the defensive lines. The automatic detection of such formations is necessary for the analysis of large datasets. Position data of the players allow for algorithm-based approaches, like template matching. Therefore, the purpose of this study was to evaluate the quality of formation recognition in handball using template matching. 20 matches of handball videos were analyzed by domain experts. The respective start, end, and defensive formation was annotated. The corresponding position data was extracted, aggregated and compared to ground truth templates. Solving the linear assignment problem, the formation with the most similar template was predicted as formation. Precision and recall were calculated and compare to the majority class baseline (assigning all formations to 6:0). A total of 1,548 formations were analyzed in the final dataset. Precision was 0.95, 0.60, and 0.52 for 6:0, 5:1, and 3:2:1, respectively. Recall was 0.97, 0.45, and 0.50 for 6:0, 5:1, and 3:2:1, respectively. Classification with template matching was therefore more accurate then the majority class baseline (accuracy = 0.85). Template matching seems to be a viable solution to classify between space-oriented (6:0) and more man-oriented (5:1, 3:2:1) formations but less able to identify differences between man-oriented formations. Future work can include more information than the average player position, e.g., the players’ covariance matrices to get more information about role specific movement behavior.

The midsole component of running shoes are an important factor for performance and user perception. The shear deformation of the midsole has been linked to smoother running patterns and is an attribute of interest to running shoe brands to optimise. A finite element (FE) model of the ground contact phase when running was created and used to simulate the shear deformation in the heel area of three sole conditions. The model was validated against ground reaction forces and all three conditions were visually inspected against high-speed video images. The model predicted different magnitudes of shear across the heel region of each sole, as well as between the different soles. Consequently, the FE model was deemed a useful tool for optimising the shear deformation of midsoles and could be used as a virtual development tool, potentially leading to more optimal solutions and reduced reliance of physical prototyping.

This systematic review aims to analyze sports injury prediction models based on machine learning technology in order to provide references for the development of sports injury prediction models. The study employed a literature review method to search the Web of Science. The models consider a range of features, including anthropometry, sports quality, training load, injury history, exercise duration, sleep, genetic information, and academic performance. Feature selection strategies applied include forward selection, Gini coefficient decline, and lasso analysis. The training algorithms utilized are support vector machines, random forest, decision trees, neural networks, and logistic regression. Model performance evaluation methods consist of HoldOut crossover and k-crossover methods, with evaluation metrics varying from AUC, sensitivity, specificity, F1 score to accuracy. Improvements are suggested to include the competition scene and sports recovery indicators in the feature extraction process. Moreover, greater diversity in training sets, utilization of unsupervised and semi-supervised learning algorithms, and implementation of multiple model optimization methods are recommended to enhance the models’ robustness and overall performance.

Researchers in computer vision are developing a method for Action Quality Assessment (AQA) that evaluates the quality of human actions in videos rather than identifying them. Specifically for figure skating, the task involves estimating the final scores from a video of a short program. It serves as an auxiliary assessment for judging skaters’ performances. Despite the significance of accurately predicting individual jump scores due to their substantial impact on final scores, prior studies have overlooked this aspect. Although videos concentrate on a solitary skater, they often include extraneous elements unrelated to assessing the quality. Consequently, expert humans discard non-essential data to make visually precise evaluations. Our research has illuminated the gaze patterns of judges and skaters when assessing jumps, developing a jump-performance prediction model that leverages their gaze patterns to filter out irrelevant information. In addition, we enhanced its predictive precision by incorporating kinematic data from a tracking system. The findings revealed a marked contrast in gaze patterns: skaters focused mainly on the face, while judges paid more attention to the lower body. Integrating these gaze patterns into our model improved its learning efficiency, with the model improved accuracy by assimilating the gaze data from both groups of specialists. Our work marks an innovative step towards merging human insight and artificial intelligence to tackle the challenge of jump performance evaluation in figure skating, offering valuable contributions to computer vision and sports science.

The aim of this study was to assess the effect of a low-cost VR goggle solution on perceived exertion, general mood, and well-being in healthy young adults. Fifteen participants completed three 25-min exercise sessions at 75% of their maximum heart rate under two conditions: with VR (using VR) and without VR (not using VR). The VR solution used a modified version of the Google LLC Cardboard and allowed for attachment of a smartphone to the head. Participants using VR were immersed in a 360º-view video simulation of an outdoor ride. Before and after the exercise sessions, participants completed PACES questionnaire to measure enjoyment, and BRUMS questionnaire to assess mood states. Perceived exertion was also measured before and after the sessions using Borg scale. Participants’ perceived exertion was higher without VR compared to exercising with VR. With VR yielded higher scores for certain positive-worded items, while exercising without VR resulted in higher scores for certain negative-worded items. When comparing mood states before and after exercising with VR, participants reported feeling less annoyed, confused, sleepy, tired, and nervous, and more active and lively. After exercising without VR, participants reported feeling more active. Our study provides evidence that affordable VR can also promote positive mood states and enhance enjoyment during exercise, thus contributing to expanding its use as an alternative to increase PA levels.

Effective utilization of tactics is paramount for achieving victory in elite table tennis. Coaches and players strive to develop tactical approaches to outperform their opponents. This work employs game theory to analyze the strategic decision-making of elite female table tennis players during matches, aiming to identify the optimal action strategies. By examining the direct scoring rate and usage rate of 11,576 tactic usage in 57 matches from top-tier tournaments in 2019, including the quarterfinals, semifinals, and finals, we reveal key insights. The results show that (1) the most rewarding serve tactic requires the player to use Pendulum serve from the middle or forehand area to the opponent’s short middle area, allowing for a Twist return to the long backhand area before executing a forceful Topspin to the opponent’s long backhand area; (2) in the receive and attack phase, the player’s optimal tactics are to use a Push to the long forehand area and a Short to the short middle area when facing server’s ball drops in short forehand and short middle area; (3) the most advantageous tactic during stalemates entails hitting the ball with Block towards the opponent’s long middle area, prompting a Topspin return to the long middle area, and subsequently using a Topspin to the opponent’s long forehand area when facing a ball in the long forehand area; (4) the optimal strategies found by game theory align with the preferences of elite female table tennis players in the use of tactics.

Performance analysis of sports using computer vision technology has been popular in recent times due to its cost efficiency, accessibility, and applicability. Despite its popularity, technological involvement is limited in its performance analysis in Taekwondo. Therefore, this study aimed to validate the usage of computer vision technology in the performance analysis of Taekwondo. In total, 5 Taekwondo players including 2 world-class and 3 master-level players were recruited for this study. They were asked to perform Taeguk Il-Chang while being captured by 12 VICON and 4 Contemplas cameras. The Contemplas cameras were calibrated using a 1m-squared calibration cage. The Contemplas recordings were processed by OpenPose to extract a 2-dimensional human skeleton which was triangulated to obtain 3-dimensional data. In total, 8 joint angles were calculated from each system and compared using the mean absolute error. The high kick during the performance was analyzed using the computed joint angles. The upper-body joint angles were more erroneous than the lower-body joint angles. The minimum hip angles and high kick completion time were lower and faster in the world-class players than the master-class players. In conclusion, the errors may be by the occlusion. The lower minimum hip angle and faster kick completion in world-class players than master-class players may mean that world-class players are more capable of kicking higher and faster. This fact can be applied to the training design for trainers and players. Despite the conclusion, the small sample size and comparability between OpenPose and a VICON system were considerable limitations.

Baseball is commonly referred to as a sport of statistics. Unlike many other sports, baseball boasts an extensive array of records that allow for comprehensive retrospection of game content and outcomes. In recent times, the use of Sabermetrics has gained prominence in baseball. Sabermetrics were developed to address the shortcomings of traditional baseball statistics. Notable metrics include batting metrics (OPS, BABIP, wOBA, IsoP) and pitching metrics (WHIP, FIP, QS%, Rel%). Numerous research studies have been conducted on Sabermetrics, including comparisons with traditional records and correlations with player salaries. However, there has been a scarcity of research focusing on which specific metrics significantly impact a team’s success. Therefore, the aim of this study is to explore the correlation between Sabermetrics and team performance in the Korean professional baseball regular season.

To achieve this research objective, we collected data on Sabermetrics batting metrics (OPS, BABIP, wOBA, IsoP), pitching metrics (WHIP, FIP, QS%, Rel%), and each team’s performance in the Korean professional baseball league from 2020 to 2022. Data were gathered from the official website of the Korean professional baseball league and the statistics website Statiz. We conducted correlation analyses using the Jamovi 2.3.26.0 software and utilized Python for data visualization.

In conclusion, this study found the following key results: Firstly, among batting metrics, IsoP and wOBA showed a close correlation with team rankings, while BABIP exhibited minimal impact on team performance. Secondly, among pitching metrics, WHIP and QS% displayed the strongest correlation with team rankings, while Rel% was found to have little relevance to team performance. Thirdly, there was a similarity in the impact of Sabermetrics and traditional metrics on team performance. This is likely due to Sabermetrics being derived from the manipulation of basic statistics.

Based on the findings of this study, improving team performance may be achieved by increasing IsoP and wOBA metrics while aiming to enhance WHIP and QS% metrics. This entails reducing opponent on-base opportunities and ensuring starting pitchers perform well over extended innings to minimize runs conceded. It is hoped that future research will explore a broader range of Sabermetrics, further enriching our understanding of their impact on team success.

Badminton is one of the most popular sports in the world. In this paper, we examined the assumption that smash in the badminton double discipline is tactically different from the single discipline and that there exists a powerful three-stroke sequence: cooperation of smash and net-kill (CoSN). Four evaluation criteria (direct scoring rate, create scoring rate, net-kill opportunity and awards) were proposed in this paper. Five smash parameters (height of impact-point, post-impact shuttlecock speed, distance from impact point to back boundary line, shuttle flight time and height of trajectory end-point) were measured and counted in a balls-into-bins model to investigate the link between scoring rate and smash parameters. We collected a dataset comprising 55,433 strokes from 46 world-class women doubles games. We found the most relevant smash parameter is shuttle flight time. The knowledge of low-cost-high-reward smash behaviour helps to improve training methods.

The FIFA World Cup is a prominent global sporting event with significant impact. This research analyzes panel data from 68 national teams that competed in the final stage of the World Cup between 1994 and 2022 (15th to 22nd editions), using the Sport Policy Factors Leading to International Sporting Success (SPLISS) model. The study seeks to identify the key systemic factors contributing to the success and winning trends in international football. The study found that (1) the country’s economic level (GDP) positively affects World Cup performance, but not significantly; (2) a brief history of the formation of national football associations (NFAHistory) and the time of national football associations joining the Intercontinental Football Confederation (IFAHistory) do not significantly contribute to football performance when they are used as a single variable; (3) the World Cup hosts (Hold), historical hosting experience (Host) and the interaction variable (Winners × CNOGMedals) will significantly affect World Cup performance (4) there is still a significant difference in football performance between different population sizes (Gpop) and different regions (Football regions) in the last eight World Cup.

In tennis match, the variability in stroke routes and landing location depth limits the opponent’s tactics and leads to winner shots. This study was aimed to explore stroking patterns of winner final shots and the hitting technique preferences of elite players when achieving such strokes. The study acquired the spatial-temporal data of all winner shots (7,126) from the 2022 Australian Open male and female singles matches. Additionally, we used clustering, kernel density estimation, and radar charts for the comprehensive statistical analysis and visual representation of stroke routes and hitting techniques. The main findings show that: (1) The strategic use of forehand inside-out/in on the backhand side was more prevalent in professional male singles matches compared to female matches; (2) Four empirical target locations for winner final shots were identified, with two in the non-deep area and two in the deep area, all positioned close to the sidelines; (3) In addition to establish baseline dominance, male players also emphasize approach and net strokes.

In badminton, achieving a return height that is higher than the net height is typically considered a better attacking advantage for winning points. However, given the limitations of methods for obtaining data, only a few studies have quantified the effect of the relationship between the distance of the hitting point from the net and the height of the hitting shuttlecock on the hitting results. In this study, we focused on analysing the complex characteristics of the combined effects of the distance between the point of the hit and the net and the height of the hit on the outcome of the men’s singles (MS) badminton player’s stroke. Results were as follows. 1) In a badminton competition, the spatial position of a player’s stroke height and the distance of the stroke point from the net limited the player’s stroke technique. 2) The distribution of wins and faults in different regions of the MS programme had its own unique characteristics in terms of the effects of hit height and distance of the point from the net on the outcome. 3) In high-level MS matches, the best scoring area was the forecourt near the net where the height of the batting point was higher than the plane of the net. The backcourt area, where the height of the hitting point was higher than the plane of the net, had the highest scoring frequency but more errors. However, the overall gain value was positive. When the height of the point was below the plane of the net, the overall scoring rate was negative, and the height of the point was not dominant.

This study extensively explores the diverse factors influencing National Basketball Association (NBA) draft rankings, utilizing a comprehensive dataset spanning from 2000 to 2022. Employing both fixed effects and random effects analyses, critical determinants affecting draft positions are revealed. Specifically, scoring efficiency, measured by Points per Point (PTS/pt) and Player Efficiency Rating (PER), demonstrates a robust correlation with higher draft rankings (β = 0.025, p = 0.023; β = 0.084, p < 0.001). Additionally, a more challenging Strength of Schedule (SoS) during a player's collegiate career positively impacts draft positions (β = 0.04, p < 0.001). Youth remains favored in the draft process, with younger draftees consistently securing higher positions (β = -0.018, p < 0.001). Background and playing experience emerge as crucial determinants of draft value. Players with a familial background in the NBA receive preferential treatment in the draft. Additionally, players from overseas leagues generally trail behind their counterparts from American collegiate leagues in NBA draft rankings. These findings shed light on the intricate landscape of NBA draft selections, providing essential insights for teams, scouts, and prospects to navigate the complex decision-making process.

The project SERVE aims to promote education in and through sports and encourage dual career of athletes. As part of the project, the best European practices in integrating digital platforms, tools, and apps into training and educational programs at European schools and sports clubs have been explored. Literature research and an online survey were carried out for this purpose. A search on European projects implemented from 2015 to 2022 regarding digitalization in sports was undertaken. Results indicated that while implementation of digitalization projects have increased in recent years, ones specifically focusing on digital technologies in sports are almost entirely absent. Digitization measures in the field of education are prevalent across Europe, but progress varies. Denmark may serve as a role model. Digital tools can enhance motivation, impart knowledge, and support motor leaning; however, they may also conflict with exercise. Interviewed participants included sports teachers, coaches, and members of volleyball clubs and federations to get a broad picture of the use of digital tools in training and sports education. Results showed differences between countries. A deficiency in the specific education that coaches and teachers need to fully exploit the technical potential of these tools is apparent. In summary, several key components of successful integration of digital tools in sports were identified.

This study investigates the existence of momentum phenomenon in recurve archery. Using data from professional recurve archery tournaments, we find archers demonstrate relatively improved shooting performance after achieving consecutive streaks of two bullseyes compared to their average scores following the first 10 points and second 9 points. This supportive evidence aligns with the notion that “success breeds success,” suggesting that a sequence of three perfect shots is a result of a hot hand rather than mere luck. Moreover, we find no gender differences in the hot hand effect. Therefore, both women’s and men’s performance is significantly affected by positive momentum.