Player Tracking Research Articles

Contextualised physical metrics: The physical demands vary with phase of play during elite soccer match play.

The physical demands of elite soccer match play have traditionally been measured using aggregated whole-match metrics. However, match play is increasingly considered as distinct phases of play, although the influence of phase of play on match physical demands remains largely unknown. This study compared physical intensity, acceleration and deceleration demands, between phases of play and according to playing position. The duration of each match from a major European league (n=1083) was divided into one of five reciprocal phases (for the in-/out-of-possession team) using event and tracking data: build-up/high-block, progression/mid-block, chance creation/low-block, fast attack/fast defence, or attacking transition/defensive transition. Player tracking data were used to calculate physical intensity as the rate of distance covered (m⋅min-1) in total and within five speed categories, and the proportion of time spent accelerating and decelerating (>2ms-2) during each phase of play. Rate of distance covered in total differed markedly with phase of play; fast attack 35%-53% greater, and fast defence 33%-50% greater, than other in-/out-of-possession phases respectively, and these effects were amplified for the rate of distance covered at higher speeds (e.g., sprinting ≥4-fold differences between phases). Match phase also affected the proportion of time spent accelerating and decelerating (highest for fast attack and chance creation, respectively), especially when in-possession for forwards and when out-of-possession for defenders (p<0.001). Phase of play had a large effect on the physical intensity of match play, particularly rates of distance covered at higher speeds, as well as the acceleration and deceleration demands, and in a position specific manner.

Empowering Student Transition and Engaging Students Using Actionbound: A Gamified Collaborative Community Learning Platform

Transition to University is a significant milestone for students who are embracing new experiences and challenges. It is particularly vital for international students who have to adapt to a new and different learning environment in the UK. Successful transition and sense of belonging in a new institution lay foundation for students’ academic success. However, transition is a process rather than an event. Accordingly, fostering a collaborative community learning environment and facilitate peer networks and community support becomes critical. It enables social interactions and active learning by engaging students in collaborative group activities, which would in turn support students’ transition to and belonging in a new learning environment. With the advancement of technology, computer-assisted and computer-supported collaborative learning flourished in 21st century. The emergence of mobile digital technology leads to the design of gamification in learning, known as game-based learning or gamified collaborative community learning in order to engage students in fun and interactive activities. ‘Actionbound’ is one of the versatile gamification platforms and mobile apps which can engage students in collaborative community learning with gamified activities. It enables users to create, share and participate in interactive scavenger hunts, tours, games and quizzes. It allows users to design custom challenges, incorporate multimedia content and track players’ progress in real-time. It engages participants to work collaboratively to solve challenges and complete tasks. At UHBS, we have used Actionbound to support the student transition into the University’ learning journey at the start of term. We used scavenger hunt game and quiz functions within Actionbound to direct students to collaborate in teams to solve the challenges and tasks. In this action research, students found Actionbound fun and loved the competitive nature with points and leaderboards. It helped them build network and form friendship. Together, they solved puzzles and tackled challenges which supported their transition and got them prepared for learning in a new environment. Moving forward, as an ongoing action research, we would like to extend the use of Actionbound to term time and in module delivery so to track students’ engagement with the collaborative learning platform and investigate the impact on their academic performance.

External Mechanical Work Versus Energy Expenditure During on-Court Tennis Drills in Elite Adolescent Tennis Players

Abstract Purpose Tennis demands exceptional technical skills, speed, strength and endurance. Previous studies have highlighted the high energy expenditure during tennis drills due to constant directional changes. While metabolic monitoring and player tracking have been used to quantify exercise intensity and movements, mechanical efficiency (ME) in tennis remains poorly understood. This study aimed to investigate ME in tennis by combining metabolic and mechanical measurements during on-court drills. Methods Ten elite adolescent tennis players (males n = 5, females n = 5) performed a VO2max test on a treadmill and participated in three standard tennis drills, Spanish cross, lateral and inside out. During the tennis drill, full-body kinematics were captured, and energy expenditure was measured using a portable breath-by-breath gas analysis. ME was calculated for each drill, and individual differences were examined. Results Results revealed significant differences in energy expenditure and mechanical work between drills. ME was not associated with VO2max or running speed. highlighting the need to focus on other factors like strength, footwork technique, and neuromuscular training to enhance ME. Conclusion This study underscores the complexity of ME in tennis, with individual variations and where factors like footwork technique, muscular strength and motor abilities might influence ME.

In-game soccer outcome prediction with offline reinforcement learning

Predicting outcomes in soccer is crucial for various stakeholders, including teams, leagues, bettors, the betting industry, media, and fans. With advancements in computer vision, player tracking data has become abundant, leading to the development of sophisticated soccer analytics models. However, existing models often rely solely on spatiotemporal features derived from player tracking data, which may not fully capture the complexities of in-game dynamics. In this paper, we present an end-to-end system that leverages raw event and tracking data to predict both offensive and defensive actions, along with the optimal decision for each game scenario, based solely on historical game data. Our model incorporates the effectiveness of these actions to accurately predict win probabilities at every minute of the game. Experimental results demonstrate the effectiveness of our approach, achieving an accuracy of 87% in predicting offensive and defensive actions. Furthermore, our in-game outcome prediction model exhibits an error rate of 0.1, outperforming counterpart models and bookmakers’ odds.

An application of clustering to classify movement patterns in men’s professional grand slam hard court tennis

ABSTRACT The movement cycles from Australian Open player tracking data were analysed using the Lloyd k-means clustering algorithm to classify movement patterns that exist in men’s grand-slam tennis. The elbow criterion method identified six movement patterns, and the k-means model allocated each movement cycle into one of these discrete groups. A description of each movement pattern is presented, outlining three inner range and three end range movement patterns, which are distinguishable by distance, direction, time pressure, and starting location. These findings provide objective details for coaches and athletes to understand tennis movement, overcoming vague descriptions of the inner range and end range in tennis vernacular. The amalgam of distance, direction, and time pressure demands that categorise the six movement patterns can enhance the specificity of training drill design and movement evaluation. Furthermore, evaluating the movement patterns a player typically elicits during match-play can inform typical load exposure and be useful in load monitoring practices. Additionally, understanding the prevalence of movement patterns in a typical match may help understand the strategic approaches players use during match-play.

An Efficient Application of Modified YOLOv5 in Basketball Player Detection and Analysis

Effective analysis helps players evaluate their performance, make necessary adjustments, and develop diverse game strategies. Moreover, the analysis provides viewers with different perspectives, enhancing their understanding of the game. This study aims to develop a basketball player detection and analysis system to assist in analyzing on-court situations. The system uses perspective transformation to obtain player tracking information on the top view image in basketball games. The system uses a modified you only look once (YOLO) v5 model that replaces the backbone of YOLOv5s with the MobileNetv3-small architecture for player detection. Compared to the original YOLOv5, the modified YOLOv5 reduces parameters from 7.02 × 106 to 3.5 × 106, a decrease of 49.8%. The number of frames obtainable per second increases from 12.4 to 17.5, an improvement of about 41.1%. Finally, the system performs perspective transformation and tracks the detected player positions onto the top-view court image using the YOLOv5 model.

High-Intensity Running During International Male Field Hockey Involves Frequent Changes of Direction and Repeated Accelerations but Seldom Reaches Sprint Velocities.

Goods, PS, Appleby, B, Scott, BR, Peeling, P, and Galna, B. High-intensity running during international male field hockey involves frequent changes of direction and repeated accelerations but seldom reaches sprint velocities. J Strength Cond Res XX(X): 000-000, 2024-The aim of this investigation was to quantify the characteristics of high-intensity running in international male field hockey. Player movement data were collected through wearable player tracking devices across 3 tournaments (17 matches) from 27 members of the Australian male field hockey team (totaling 266 player matches). Active duration (minute), high-intensity efforts (>2.5 m·s-2 for >1 second), repeated high-intensity efforts (≥3 efforts with ≤45-second recovery between efforts), and sprints (>7 m·s-1) were extracted and aggregated for each player match. The duration, distance, mean and peak speed, mean change of direction and straightness for each high-intensity effort, and peak deceleration at the conclusion of each effort were calculated. Mixed-effects models were used to estimate the mean for each outcome (fixed effect), with random intercepts modeled for player and match. On average, players completed 42.2 high-intensity efforts and 2.1 sprints per match. High-intensity efforts were short (3.61 seconds, 12.9 m), failed to reach high velocities (average peak velocity: 4.8 m·s-1), and involved frequent changes of direction (56% of efforts were not straight). There were 4.7 bouts of repeated high-intensity efforts per player, per match (comprising 3.8 efforts, with 3.53 seconds per effort separated by 16.2-second recovery). Last, we also found that 29% of high-intensity efforts involved a high-speed deceleration (>3.5 m·s-2), and these occurred frequently when in conjunction with a >45° (86%) or >90° (89%) change of direction. We recommend practitioners focus on developing players' ability to repeatedly accelerate, decelerate, and change direction over short distances and adopt testing and monitoring programs that assess these qualities.

A Review of the Validity and Reliability of Accelerometer-Based Metrics From Upper Back-Mounted GNSS Player Tracking Systems for Athlete Training Load Monitoring.

Dawson, L, Beato, M, Devereux, G, and McErlain-Naylor, SA. A review of the validity and reliability of accelerometer-based metrics from upper back-mounted GNSS player tracking systems for athlete training load monitoring. J Strength Cond Res 38(8): e459-e474, 2024-Athlete load monitoring using upper back-mounted global navigation satellite system (GNSS) player tracking is common within many team sports. However, accelerometer-based load monitoring may provide information that cannot be achieved with GNSS alone. This review focuses on the accelerometer-based metrics quantifying the accumulation of accelerations as an estimation of athlete training load, appraising the validity and reliability of accelerometer use in upper back-mounted GNSS player tracking systems, the accelerometer-based metrics, and their potential for application within athlete monitoring. Reliability of GNSS-housed accelerometers and accelerometer-based metrics are dependent on the equipment model, signal processing methods, and the activity being monitored. Furthermore, GNSS unit placement on the upper back may be suboptimal for accelerometer-based estimation of mechanical load. Because there are currently no feasible gold standard comparisons for field-based whole-body biomechanical load, the validity of accelerometer-based load metrics has largely been considered in relation to other measures of training load and exercise intensity. In terms of convergent validity, accelerometer-based metrics (e.g., PlayerLoad, Dynamic Stress Load, Body Load) have correlated, albeit with varying magnitudes and certainty, with measures of internal physiological load, exercise intensity, total distance, collisions and impacts, fatigue, and injury risk and incidence. Currently, comparisons of these metrics should not be made between athletes because of mass or technique differences or between manufacturers because of processing variations. Notable areas for further study include the associations between accelerometer-based metrics and other parts of biomechanical load-adaptation pathways of interest, such as internal biomechanical loads or methods of manipulating these metrics through effective training design.

Multi Player Tracking in Ice Hockey with Homographic Projections

Multi Player Tracking in Ice Hockey with Homographic Projections

Is Thomas Tuchel wrong? Evaluation of hexagonal shaped drills based on machine learning and position data

As a variation of small-sided games (SSGs), Thomas Tuchel's hexagonal shaped possession drills attracted a lot of attention in coaches’ soccer education. Changing certain variables (e.g. pitch size) is one traditional approach to provide an optimal stimulus for a specific training goal. Therefore, the presented field study investigated whether hexagonal shaped drills show changes in tactical key performance indicators (KPIs) using positional data in a controlled experimental setting. Data were collected using player tracking systems (1 Hz) in 5 versus 5 SSGs in both the full-size pitch and pitch with cut-off corners condition. At match-related level, trial duration and outcome were examined. At player-related level, the tactical KPIs effective playing space (EPS), length-per-width ratio, space control and overplayed defenders were analyzed. The results show significant differences in length-per-width ratio (attacking team), the EPS and space control (30m-zone). However, it could not be confirmed that hexagonal shaped drills are played more vertically and faster toward the goal than on a normal pitch, as there were no changes in trial duration, outcome or overplayed defenders. Nevertheless, an experimental positional data analysis paradigm is a useful approach to investigate tactical principles in high-level professional soccer.

Soccer match broadcast video analysis method based on detection and tracking

AbstractWe propose a comprehensive soccer match video analysis pipeline tailored for broadcast footage, which encompasses three pivotal stages: soccer field localization, player tracking, and soccer ball detection. Firstly, we introduce sports camera calibration to seamlessly map soccer field images from match videos onto a standardized two‐dimensional soccer field template. This addresses the challenge of consistent analysis across video frames amid continuous camera angle changes. Secondly, given challenges such as occlusions, high‐speed movements, and dynamic camera perspectives, obtaining accurate position data for players and the soccer ball is non‐trivial. To mitigate this, we curate a large‐scale, high‐precision soccer ball detection dataset and devise a robust detection model, which achieved the of 80.9%. Additionally, we develop a high‐speed, efficient, and lightweight tracking model to ensure precise player tracking. Through the integration of these modules, our pipeline focuses on real‐time analysis of the current camera lens content during matches, facilitating rapid and accurate computation and analysis while offering intuitive visualizations.

Analyzing Players' Tactical Positions and Movement Trajectories in Soccer Matches Using Machine Vision Algorithms

Tactical positions and movement trajectories in soccer matches using machine vision algorithms involves leveraging computer vision techniques to track players and the ball throughout the game. By processing video feeds from multiple cameras, machine vision algorithms can identify players, recognize their positions on the field, and map out their movement trajectories over time. This analysis provides valuable insights into team formations, player positioning, and tactical strategies employed during matches. Coaches and analysts can use this information to assess team performance, identify patterns in gameplay, and make data-driven decisions to optimize tactics and training regimes. Machine vision algorithms enhance the understanding of soccer dynamics, facilitating more strategic and effective approaches to coaching and gameplay.. This paper presents a comprehensive analysis of player movement, tactical positions, and detection in soccer matches using machine learning algorithms. Leveraging a simulation environment built upon a Recurrent Neural Network (RNN) model trained on historical match data, we investigate the efficacy of the model in accurately predicting future player positions based on contextual features such as player trajectories and ball movement. Furthermore, we explore the model's ability to estimate tactical movements during specific time intervals, providing valuable insights for coaches and analysts in understanding team strategies and adapting to match dynamics. Additionally, we evaluate player detection systems to assess their capabilities and limitations, laying the groundwork for future improvements in player tracking technologies. Our findings underscore the potential of machine learning in soccer analytics, offering actionable insights for enhancing performance analysis, strategic decision-making, and overall understanding of the game. Our results show a high prediction accuracy, with an average Mean Squared Error (MSE) of 0.05. Furthermore, we explore the model's ability to estimate tactical movements during specific time intervals, achieving an overall prediction accuracy of 85%. Additionally, we evaluate player detection systems to assess their capabilities and limitations, achieving a detection accuracy of 90%.

Spatial roles in hockey special teams

Abstract Special teams (i.e. power play and penalty kill) situations play an outsized role in determining the outcome of ice hockey games. Yet, quantitative methods for characterizing special teams tactics are limited. This work focuses on team structure and player deployment during in-zone special teams possessions. Leveraging player and puck tracking data from the National Hockey League (NHL), a framework is developed for describing player positioning during 5-on-4 power play and 4-on-5 penalty kill possessions. More specifically, player roles are defined directly from the player tracking data using non-negative matrix factorization, and every player is allocated a unique role at every frame of tracking data by solving a linear assignment problem. Team formations naturally arise through the combination of roles occupied in a frame. Roles that vary on a per-frame basis allow for a fine-grained analysis of team structure. This property of the roles-based representation is used to group together similar power play possessions using latent Dirichlet allocation, a topic modelling technique. The concept of assignments, which remain constant over an entire possession, is also introduced. Assignments provide a more stable measure of player positioning, which may be preferable when assessing deployment over longer periods of time.

Strengthening oversight and integrity: The multi-faceted role of centralized player tracking systems in gambling.

Strengthening oversight and integrity: The multi-faceted role of centralized player tracking systems in gambling.

Soccer player tracking and data correction based on attention with full-field videos

Soccer player tracking and data correction based on attention with full-field videos

Player tracking in itself may not be enough for efficient reduction of gambling harm, but combined with a personal gambling license could provide new possibilities to tracking gambling-related harm.

Player tracking in itself may not be enough for efficient reduction of gambling harm, but combined with a personal gambling license could provide new possibilities to tracking gambling-related harm.

Beyond 'single customer view': Player tracking's potential role in understanding and reducing gambling-related harm.

Usage of electronic gaming machines (EGMs) and on-line gambling is strongly associated with gambling-related harm. Player-tracking systems can monitor a gambler's activity across multiple sessions and/or operators, providing a clearer picture of the person's risk of harm with respect to these gambling formats and enabling harm reduction efforts. The Finnish and Norwegian state monopolies have player-tracking systems in place, while the United Kingdom is implementing an operator-led system called 'single customer view' for on-line gambling, and Australian states are proposing similar 'player cards' for land-based EGMs. Player tracking can advance harm reduction efforts in three ways. First, player tracking improves our understanding of gambling-related harm by providing data on how the population gambles, which can potentially be linked with operator, government and/or prevalence data sets. Secondly, player tracking can be used to implement harm reduction measures such as expenditure limits, self-exclusion and age verification. Thirdly, player tracking can provide a platform to evaluate harm reduction measures via gold-standard field trials. These potential benefits need to be weighed against various practical and ethical issues. The potential benefits of player-tracking systems would be maximized via systems administered independently of the gambling industry and implemented universally across all gambling in a given jurisdiction.

Football Player Tracking and Performance Analysis Using the OpenCV Library

Football Player Tracking and Performance Analysis Using the OpenCV Library

Advanced Active Player Tracking System in Handball Videos Using Multi-Deep Sort Algorithm with GAN Approach

Advanced Active Player Tracking System in Handball Videos Using Multi-Deep Sort Algorithm with GAN Approach

Optimizing Semantic Segmentation for Enhanced Football Analytics: A Pixel-level Approach

Semantic Segmentation, a pivotal technique in image analysis, is adeptly leveraged in this research to bolster sports analytics, with a concentrated focus on football. A comprehensive pipeline is unveiled for an in-depth analysis of a select portion of the IAUFD 100k dataset, encompassing 2030 manually annotated football images. The methodology entails a thorough evaluation and comparison of diverse semantic segmentation models, supplemented by the integration of advanced pre-processing strategies and optimal training techniques. Such a holistic approach culminates in a marked enhancement in model performance, as evidenced by a significant uptick in the mean Intersection over Union (mIoU). This research offers granular, object-oriented insights that substantially augment player tracking, action recognition, and event detection in football. The conclusive remarks of the study highlight prospective avenues for further research, emphasizing the potential incorporation of Explainable AI and advanced Metamorphic and Security Testing to fortify sports analytics.