⚽️ Markerless Motion Analysis of Football Goalkeepers During Penalty Kicks from Broadcast Videos
2024–2025 · Jaka Kuraj, Idsart Kingma, Mauricio Verano Merino, and Rony Ibrahim
To examine the relationship between goalkeepers’ preparatory posture, stance width after sidestepping, and timing variables with penalty save success in real games.
Outcomes: No significant predictors of saves were found, but goalkeepers adopted a wider preparatory stance than in lab settings and typically decreased stance width after sidestepping, contrasting with laboratory findings.
🏑 Automated Player Tracking in Field Hockey Baseline Comparison
2024–2025 · Ezra Eijsenring, Mauricio Verano Merino
Design and evaluate an end-to-end multi-camera player tracking system for professional field hockey, establishing a performance baseline for automated tracking in the sport.
Outcomes: Developed a quantitative benchmark for field hockey player tracking, achieving a mean positional error of 1.94 m, RMS error of 2.07 m, and an F1-score of 5.0%, providing a foundation for future model specialization and accuracy improvements.
⚽️ Computer Vision-Based Automation of Football Performance Analytics Detecting Player Body Orientation and Scanning Behavior
2024–2025 · Sobhaan ul Husan, Mauricio Verano Merino, and Eli Sarasola
To develop an automated computer vision pipeline for estimating player body orientation and pre-reception scanning behavior in football, reducing the need for manual video annotation.
Outcomes: Achieved 75% accuracy in body orientation detection and established a foundational framework for automated scan counting, demonstrating the potential to significantly reduce manual analysis time in football performance analytics despite video resolution constraints.
🏋️♀️ Classification of Movement Phases in Olympic Weightlifting Using Inertial Measurement Units
2024–2025 · Sam Hartogs and Anil Yaman
To develop an automated classification system for the sub-phases of the snatch and clean and jerk (C&J) using Inertial Measurement Units (IMUs) and machine learning algorithms.
Outcomes: The provision of accurate, phase-specific performance metrics that enable coaches to mitigate injury risks and optimize athlete technique through data-driven feedback and longitudinal progression monitoring.
⚽️ Paying for Popularity The Value of Online Attention in Soccer
2024–2025 · Vivian Witting, Mauricio Verano Merino
To analyze the relationship between online attention and the market values of professional soccer players using multi-platform sentiment and engagement data.
Outcomes: Demonstrated that online attention metrics improved market value prediction accuracy by up to 18% (R² increase), with peaks in mentions and positive sentiment significantly correlating with subsequent value rises, revealing the predictive and explanatory power of social media and news dynamics in player valuation.
⚽️ Transfer Learning for Semi-Supervised Football Event Classification Using Tracking Data
2024–2025 · Minh Duc Nguyen, Mauricio Verano Merino
To develop a semi-supervised transfer learning framework for accurate classification of football on-pitch events from scarce and imbalanced tracking data.
Outcomes: Improved event classification performance with increases of up to 15% in F1-score, 12% in precision, and 10% in AUC across both majority and minority classes by integrating latent-space oversampling and domain-adapted pretraining on high- and low-quality tracking datasets.
🥎 Machine and deep learning models for batted ball velocity prediction in virtual environments
2023–2024 · Lloyd Nyarko, Daniel Müller, Mauricio Verano Merino, David Mann
To enhance a baseball batting virtual reality training environment by predicting batted ball velocities from bat swing data using machine learning and deep learning models.
Outcomes: Developed and integrated Bayesian ridge regression and multi-layer perceptron models that achieved comparable predictive accuracy to a physics-based heuristic baseline, validating the feasibility of AI-driven ball-bat collision modeling for improving realism and feedback in VR-based baseball training systems.
🏋️♀️ Accessible Performance Analytics for Olympic Weightlifting based on Computer Vision
2023–2024 · Freek Cool and Anil Yaman
To design and validate a computer vision-based framework capable of extracting kinematic performance statistics from standard video recordings of Olympic weightlifting movements.
Outcomes: A low-barrier, sensor-free analysis tool that provides athletes and coaches with objective performance data, removing the need for specialized hardware while maintaining the ability to monitor technique and progress.
