Video-based Tracking System Research Articles

Interpersonal coordination of opposing player dyads during attacks performed in official football matches

ABSTRACT The purpose of this study was to characterise the interpersonal coordination between opponent players during offensive sequences in official matches and to verify if offensive sequences ended in shots to goal present different coordination patterns when compared than those that ended in defensive tackles. A total of 580 offensive sequences occurred during matches resulting in shots to goal (n = 172) or defensive tackles (n = 408) were analysed. The bidimensional coordinates and technical actions of male professional football players (n = 1160) were obtained using a video-based tracking system. Dyads were defined using a network analysis and composed of the nearest opponent. Interpersonal coordination of the dyads was analysed using the vector coding and the frequency for each coordination pattern was computed. In-phase was predominant for all displacement directions and offensive sequences outcomes, and antiphase was the least frequent. For lateral displacements, offensive sequences ending in shot to goal presented lower frequency for in-phase and higher frequency for offensive player phase than ended in defensive tackle. This information about the relationship of opponent players dyads during decisive moments of the matches provides fundamentals for future research and assists coaches to understand the different behaviours in successful and unsuccessful attacks.

Investigating the Knuckleball Effect in Soccer Using a Smart Ball and Training Machine.

The term knuckleball in sporting jargon is used to describe a ball that has been launched with minimal spin, resulting in a trajectory that is erratic and unpredictable. This phenomenon was first observed in baseball (where the term originated) and has since been observed in other sports. While knuckleball has long fascinated the scientific community, the bulk of research has primarily focused on knuckleball as it occurs in baseball. Following the changes in the design of the soccer ball after the 2006 World Cup, knuckleball and ball aerodynamics were exploited by soccer players. This research examined the properties of a knuckleball in the sport of soccer. We designed and evaluated a system that could reproduce the knuckleball effect on soccer balls based on previous theories and characteristics outlined in our literature review. Our system is comprised of the Adidas miCoach Smart Ball, a companion smart phone app for data collection, a ball-launching machine with programmable functions, and a video-based tracking system and Tracker motion analysis software. The results from the testing showed that our system was successfully able to produce knuckleball behaviour on the football in a highly consistent manner. This verified the dynamic models of knuckleball that we outline. While a small portion of the data showed some lateral deviations (zig-zag trajectory), this erratic and unpredictable trajectory was much smaller in magnitude when compared to examples seen in professional games. The sensor data from the miCoach app and trajectory data from the Tracker motion analysis software, showed that the knuckleballs were consistently reproduced in-line with theoretical dynamics.

Drone-Based Position Detection in Sports-Validation and Applications.

Radio and video-based electronic performance and tracking systems (EPTS) for position detection are widely used in a variety of sports. In this paper, the authors introduce an innovative approach to video-based tracking that uses a single camera attached to a drone to capture an area of interest from a bird’s eye view. This pilot validation study showcases several applications of this novel approach for the analysis of game and racket sports. To this end, the authors compared positional data retrieved from video footage recorded using a drone with positional data obtained from established radio-based systems in three different setups: a tennis match during training with the drone hovering at a height of 27 m, a small-sided soccer game with the drone at a height of 50 m, and an Ultimate Frisbee match with the drone at a height of 85 m. For each type of playing surface, clay (tennis) and grass (soccer and Ultimate), the drone-based system demonstrated acceptable static accuracy with root mean square errors of 0.02 m (clay) and 0.15 m (grass). The total distance measured using the drone-based system showed an absolute difference of 2.78% in Ultimate and 2.36% in soccer, when compared to an established GPS system and an absolute difference of 2.68% in tennis, when compared to a state-of-the-art LPS. The overall ICC value for consistency was 0.998. Further applications of a drone-based EPTS and the collected positional data in the context of performance analysis are discussed. Based on the findings of this pilot validation study, we conclude that drone-based position detection could serve as a promising alternative to existing EPTS but would benefit from further comparisons in dynamic settings and across different sports.

Wireless AI-Powered IoT Sensors for Laboratory Mice Behavior Recognition

More than 100 million animals are used in research, education, and testing per year, and 95% of them are mice and rats. We have developed wireless artificial intelligent (AI)-powered Internet of Things (IoT) sensors (AIIS) for laboratory mice motion recognition utilizing embedded microinertial measurement units (uIMUs)—a new sensing platform fills an important research gap of monitoring behaviors of many laboratory mice in parallel. We have demonstrated a wireless IoT sensor that could be attached and carried by mice (i.e., animals that typically weigh only 20 g and with body length of ~10 cm) and used the collected motion data to recognize five common mice behaviors (e.g., sleeping, walking, rearing, digging, and shaking) in cages with an accuracy of ~76%. For comparison, current commercial video-based tracking systems that track animal behaviors in real time can reach only 70% accuracy and with limited number of parallelly tracked animals. Furthermore, several machine learning algorithms were explored to solve the imbalanced sample data problem, which allowed the accuracy of mice motion recognition to improve from ~48% to ~76% (if shaking is removed from classification, an average accuracy of 86.46% could be achieved). Less frequent mice behaviors, such as rearing, digging, grooming, drinking, and scratching, could also be recognized at an average accuracy of 96.35%. We believe this work has the potential to revolutionize animal behavioral tracking methodology by offering a solution for large batches of simultaneous small animal motion tracking and AI-based behavior recognition.

The influence of the futsal outfield goalkeeper on players' running performance

PurposeThis study aimed to investigate the influence of the outfield goalkeeper game situation on the running performance of teammates and opponent players during professional futsal matches.MethodsThe movement trajectories of 126 professional players during 6 official matches were analysed with a videobased tracking system (30 Hz). Distance covered and percentage of distance covered in different speed ranges were calculated to identify differences per team and per sub-phase of the game (traditional vs. outfield goalkeeper situation).ResultsThe teammates of the outfield goalkeeper spent a higher percentage of the distance covered in the standing and walking speed range (36.27 ± 15.38%) compared with the teammates of the traditional goalkeeper (31.44 ± 7.74%). The outfield goalkeeper spent a smaller percentage of the distance covered in the standing and walking range (41.13 ± 10.86% vs. 80.77 ± 4.30%) and a higher percentage in the low- and medium-intensity running range (31.59 ± 5.59% vs. 15.07 ± 2.74% and 14.33 ± 3.90% vs. 2.50 ± 1.19%, respectively), as well as covered a greater distance (77.41 ± 16.00 m/min) than the traditional goalkeeper (33.81 ± 3.66 m/min).ConclusionsThese findings provide coaches with a background concerning the requirements during the outfield goalkeeper situation as guidelines for final exercises in training sessions.

Football player dominant region determined by a novel model based on instantaneous kinematics variables

Dominant regions are defined as regions of the pitch where a player can reach before any other and are commonly determined without considering the free-spaces in the pitch. We presented an approach to football players’ dominant regions analysis, based on movement models created from players’ positions, displacement, velocity, and acceleration vectors. 109 Brazilian male professional football players were analysed during official matches, computing over 15 million positional data obtained by video-based tracking system. Movement models were created based on players’ instantaneous vectorial kinematics variables, then probabilities models and dominant regions were determined. Accuracy in determining dominant regions by the proposed model was tested for different time-lag windows. We calculated the areas of dominant, free-spaces, and Voronoi regions. Mean correct predictions of dominant region were 96.56%, 88.64%, and 72.31% for one, two, and three seconds, respectively. Dominant regions areas were lower than the ones computed by Voronoi, with median values of 73 and 171 m2, respectively. A median value of 5537 m2 was presented for free-space regions, representing a large part of the pitch. The proposed movement model proved to be more realistic, representing the match dynamics and can be a useful method to evaluate the players’ tactical behaviours during matches.

Interpersonal interaction during official soccer matches considering the coupling of different playing positions

ABSTRACT We analysed the interpersonal interaction between players during official matches considering the coupling of different playing positions. 2D coordinates and technical actions of 204 male professional football players during eight matches were obtained using a video-based tracking system. Dyads were composed of the nearest opponents and the interpersonal distance and frequency were computed. Dyadic distances were used to estimate the pressure index of defensive teams and the relationship with technical performance. Full-backs presented the greatest distances of their opponents. Dyads between defenders and forwards presented the shortest distances and were more frequent. In moments with ball possession the midfielders and forwards have more than one defensive player as the nearest opponent to them. The pressure index of defensive teams was significantly negatively and positively correlated with tackles performed by them, and successful passes of offensive teams, respectively, suggesting that defensive players more distant from the offensive players performed less tackles and provided better conditions for their opponents to perform passes. Interpersonal interaction is influenced by playing positions and maintaining proximity to the opponents can improve defensive technical performance. Information about the interpersonal interactions in real match situations provides the fundamentals for future research and assists coaches in planning tactical training.

A non-human primate model for stable chronic Parkinson’s disease induced by MPTP administration based on individual behavioral quantification

BackgroundThe guidelines for applying individual adjustments to macaques according to the severity of behavioral symptoms during 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) treatment were provided to reproduce stable chronic Parkinsonism in a recent study (Potts et al., 2014). But, since there are insufficient guidelines regarding objective severity criteria of individual symptoms for adjustments of MPTP treatment, it is difficult to develop MPTP-induced chronic non-human primate (NHP) models with stable symptoms. New methodThe individual adjustments of MPTP administration based on results of automatic quantification of global activity (GA) using a video-based tracking system were applied to develop MPTP-PD model. Low-dose (0.2 mg/kg) intramuscular injection was repeated continuously until GA was lower than 8% of baseline Parkinsonian behavior scores. The positron emission tomography imaging were used to follow the longitudinal course of Parkinson’s disease (PD). ResultsSignificant reductions in GA and dopamine transporter activity, along with significant increases in Parkinsonian behavior scores were found from 4 to 48 weeks following the first administration. GA was correlated with the Parkinsonian behavior score. The dopamine transporter activity was correlated with GA and the Parkinsonian behavior score. However, it was not correlated with the total dose of MPTP. Damage of dopaminergic neuronal systems in the basal ganglia was confirmed by immunohistochemistry and Western blot. Comparison with existing methodThis study reinforces previous guidelines regarding production of NHP models with stable Parkinsonian symptoms. ConclusionsThis novel strategy of MPTP administration based on global activity evaluations provides an important conceptual advance for the development of chronic NHP Parkinsonian models.

Interpersonal coordination analysis of tennis players from different levels during official matches

The purpose of this study was to assess the interpersonal coordination during official male tennis matches in players of different skill levels. Players’ trajectories of three levels (Juvenile, ATP-Future, ATP-250) were obtained using video-based tracking system. A vector coding technique was applied to obtain players’ interpersonal coordination in four coordination patterns: “anti-phase”, “in-phase”, “serving player phase” and “returning player phase”. These patterns allowed identification of the nature of the coupling and lead/lag relations between players. In all categories, players presented higher degree of “anti-phase” and “serving player phase” (when only the serving player is moving on the court or his opponent is moving with a time lag) coordination. Young players spent more time in “serving player phase” during lateral displacements than professional players. On the other hand, professional players spent more time in “returning player phase” (when only the returning player is moving on the court or his opponent is moving with a time lag) during antero posterior displacements, than young players. Interpersonal coordination did not change from the first to the second set of the match, showing that tennis players maintain their displacement characteristics and strategy, independently of proficiency level. The vector coding technique allowed to identify new coordination patterns in tennis, providing additional information about tennis dynamics and how players from different categories and proficiency levels behave during the matches.

The efficacy of a video-based marker-less tracking system for gait analysis

Most clinical gait analyses are conducted using motion capture systems which track retro-reflective markers that are placed on key landmarks of the participants. An alternative to a three-dimensional (3D) motion capture, marker-based, optical camera system may be a marker-less video-based tracking system. The aim of our study was to investigate the efficacy of the use of a marker-less tracking system in the calculation of 3D joint angles for possible use in clinical gait analysis. Ten participants walked and jogged on a treadmill and their kinematic data were captured with a marker and marker-less tracking system simultaneously. The hip, knee and ankle angles in the frontal, sagittal and transverse planes were computed. Root Mean Square differences (RMSdiff) between corresponding angles for each participant’s support phase were calculated and averaged to derive the mean within-subject RMSdiff. These within-subject means were averaged to obtain the mean between-subject RMSdiff for the relevant joint angles in the two gait conditions (walking and jogging). The RMSdiff between the two tracking systems was less than 1° for all rotations of the three joint angles of the hip and knee. However, there were slightly larger differences in the ankle joint angles. The results of this study suggest a potential application in gait analysis in clinical settings where observations of anatomical motions may provide meaningful feedback.

Coordination analysis of players’ distribution in football using cross-correlation and vector coding techniques

ABSTRACTThe purpose of this study was to investigate the coordination between teams spread during football matches using cross-correlation and vector coding techniques. Using a video-based tracking system, we obtained the trajectories of 257 players during 10 matches. Team spread was calculated as functions of time. For a general coordination description, we calculated the cross-correlation between the signals. Vector coding was used to identify the coordination patterns between teams during offensive sequences that ended in shots on goal or defensive tackles. Cross-correlation showed that opponent teams have a tendency to present in-phase coordination, with a short time lag. During offensive sequences, vector coding results showed that, although in-phase coordination dominated, other patterns were observed. We verified that during the early stages, offensive sequences ending in shots on goal present greater anti-phase and attacking team phase periods, compared to sequences ending in tackles. Results suggest that the attacking team may seek to present a contrary behaviour of its opponent (or may lead the adversary behaviour) in the beginning of the attacking play, regarding to the distribution strategy, to increase the chances of a shot on goal. The techniques allowed detecting the coordination patterns between teams, providing additional information about football dynamics and players’ interaction.

Construct validity of a video-tracking system based on orthogonal cameras approach for objective assessment of laparoscopic skills

This study was aimed to establish the construct validity of a video-tracking system based on orthogonal cameras approach for assessment of laparoscopic psychomotor skills in training environments. The camera-tracking system consists of two webcams placed in orthogonal configuration at a distance of 13.5 cm. The orthogonal cameras employ a color segmentation algorithm to register the 3D motions of the laparoscopic instruments using colored tapes placed on the distal end. For construct validity, 31 participants (4 experts and 27 residents) performed three training tasks in a laparoscopic box trainer with the built-in orthogonal cameras system. Eleven motion-related parameters were used to evaluate their performance. Statistical analysis was performed, and results between two groups were compared using a Mann-Whitney U-test. Construct validity results showed statistical differences in almost all motion-related parameters for assessment of laparoscopic technical skills. Results demonstrated that the orthogonal video-based tracking system was able to differentiate laparoscopic experience between experts and trainees surgeons. The orthogonal cameras system was successfully validated in a laparoscopic box trainer. This video-based tracking system was able to distinguish performance between experts and trainees surgeons, showing its potential as a reliable tool to assess laparoscopic psychomotor skills. The orthogonal cameras allow incorporating the advantages of this video motion-tracking technology with the benefits of the traditional laparoscopic box trainers, creating realistic haptic feedback and allowing the evaluation of psychomotor skills of the surgeons.

EP-1495: Validation of a video-based tumor tracking system for brain stereotactic cancer patients

EP-1495: Validation of a video-based tumor tracking system for brain stereotactic cancer patients

Video-based Gait Analysis for Functional Evaluation of Healing Achilles Tendon in Rats

Video-based walking track systems have been developed for gait analysis in rat models. However, there is no previous study using video-based tracking systems to address the gait parameters to evaluate the recovery of Achilles tendon rupture models. This study conducted a comprehensive gait analysis using a video-based image processing system. Eighteen Sprague-Dawley rats were randomly assigned to one of three interventional conditions: sham surgery, Achilles tendon repair, and Achilles tendon defect. After surgery, all animals were evaluated using a video-based walking track system. The gait parameters and the Achilles functional index (AFI) were further analyzed. The ankle joint angles of the injury side at mid-stance and pre-swing were highly correlated with the AFI. However, lack of sensitivity was found for the AFI. Increased measurement sensitivity of the Achilles tendon healing condition was found in the ankle joint angle of the involved side at the pre-swing and the level of asymmetry of the hindlimb joint position and stance/swing time. The overall sensitivity of the ankle motion analysis was significantly higher than that of AFI. We conclude that the ankle motion analysis is a reliable, reproducible, and sensitive tool for Achilles tendon analysis in rats.

Automation of the free-exploratory paradigm

The free-exploratory paradigm (FEP) is currently the only proposed animal model of trait anxiety, making it highly valuable to behavioural neuroscience. However, FEP has not yet been automated, so its results depend on human scoring, which can be quite imprecise. The aim of this study was, therefore, to validate an automated version of FEP, using a commercially available video-tracking system (ANY-maze© – Stoelting Co., USA). To achieve this, two experiments were performed. The first one evaluated the reliability of the video-based automation of FEP, and the second, assessed whether the zeolites, used as a bedding material in the first experiment to facilitate video-tracking, influenced the animals’ behaviour in FEP. In experiment I, 15 drug-naive, adult, male rats were tested in FEP, while their behaviour was simultaneously evaluated by ANY-maze© and two human observers. Subsequently, the intraclass correlation coefficient (ICC) was calculated for the automated and manual results of the parameters “percentage of time in the novel side” (%TNS) and “total units visited” (TUV). The analysis resulted in high, significant values of ICC (%TNS: 0.9962 and TUV: 0.9453). In experiment II, 18 drug-naive, adult, male rats were allocated to two different groups: (1) tested in FEP with zeolites; and (2) tested in FEP with sawdust. The data obtained were analysed using the Student's t-test, which revealed no significant differences between the groups for the parameters %TNS and TUV. In conclusion, the data presented here show that automation of FEP, using a video-based tracking system, is not only possible, but also highly reliable.

Shared-memory parallel multi-target tracking

The application of particle filtering in real video-based multi-target tracking systems is limited because of its high computational complexity.To overcome the efficiency bottleneck,a coarse-grained parallel multi-target tracking implementation based on the OpenMP-specified shared-memory parallel programming model was explored.A list of tracked targets was maintained in a shared variable,and each target was tracked by an independent particle filter.The number of threads and the number of targets tracked by each thread were determined by the number of processing units.Compared to its corresponding optimized sequential version,the parallel implementation,which increases the number of targets in real-time tracking from 2 to 8,is of much more practical value.

Real-time video stabilisation

This paper deals with image stabilisation for video-based tracking systems. An introduction to image stabilisation is stated. A short description of known algorithms for image stabilisation follows, including our solution based on these methods with some optimisations. At the end, we represent a suitable hardware platform, which we developed and constructed and uses Digital Signal Processor (DSP), Field Programmable Gate Array (FPGA) and Synchronous Dynamic Random Access Memory (SDRAM). This device is under testing now.

Functional Outcomes of Cleft Lip Surgery. Part II: Quantification of Nasolabial Movement

To explore nasolabial movements in participants with repaired cleft lip and palate. A parallel, three-group, nonrandomized clinical trial. Group 1=31 participants with a cleft lip slated for revision surgery (revision), group 2=32 participants with a cleft lip who did not have surgery (nonrevision), and group 3=37 noncleft control participants. Three-dimensional movements were assessed using a video-based tracking system that captured movement of 38 landmarks placed at specific sites on the face during instructed maximum smile, cheek puff, lip purse, mouth opening, and natural smile. Measurements were made at two time points at least 1 week and no greater than 3 months apart. Summary measurements were generated for the magnitude of upper lip, lower lip, and lower jaw movements and the asymmetry of upper lip movement. Separate regression models were fitted to each of the summary measurements. Lateral movements of the upper lip were greater than vertical movements. Relative to the noncleft group, the revision and nonrevision groups demonstrated 6% to 28% less upper lip movements, with the smiles having the most restriction in movement and greater asymmetry of upper lip movement. Having an alveolar bone graft further increased the asymmetry, while a bilateral cleft lip decreased the asymmetry. Lower jaw movement caused a small increase in upper lip movement. The objective measurement of movement may be used as an outcome measure for cleft lip surgery.

Patients with hemianopic alexia adopt an inefficient eye movement strategy when reading text

Patients with an acquired homonymous hemianopia often adapt over a period of a few months to compensate for some of the impairments caused by their visual field defect. Changes in their eye movement patterns have been demonstrated as performance on visual tasks improves with time; however, these patients often complain of persistent text reading problems. Using a video-based eye-movement tracking system, we investigated the text reading behaviour of patients with established hemianopic alexia (>6 months post deficit), a condition affecting left-to-right readers, with a homonymous field defect that encroaches into their right foveal/parafoveal visual field. Word-based analyses of text reading are standard in experiments involving normal readers, but this is the first time these methods have been extended to patients with hemianopic alexia. Using this method, we compared the patients' reading scanpaths to those generated by normal controls reading the same passages, and a random model generated by matching the patients' eye movement data to random permutations of the text they read. We demonstrate that patients adopt an inefficient reading strategy, fixating to the left of the preferred viewing location of words of four letters and longer. Fixating to the left of the normal preferred viewing location not only results in less of the fixated word being processed by the language system; ensuing fixations are also more likely to land within the same word (a refixation). It is this refixation rate that is the main factor in slowing reading times in these patients. Our data suggests that patients are able to extract some useful visual information from text to aid the planning of reading scanpaths as their behaviour differs critically from the random model. Potential reasons for this patient group failing to produce an effective reading strategy are discussed.

Finger movements during reach-to-grasp in the monkey: amplitude scaling of a temporal synergy

To reduce the complexity of controlling hand-shaping, recent evidence suggests that the central nervous system uses synergies. In this study, two Rhesus monkeys reached-to-grasp 15 objects, varying in geometric properties, at five grasp force levels. Hand kinematics were recorded using a video-based tracking system. Individual finger movements were described as vectors varying in length and angle. Inflection points (i.e., stereotypic minima/maxima in the temporal profile of each finger vector) exhibited a temporal synchrony for individual fingers and in the coupling across fingers. Inflection point amplitudes varied significantly across objects grasped, scaling linearly with the object grasp dimension. Thus, differences in the vectors as a function of the objects were in the relative scaling of the vector parameters over time rather than a change in the temporal structure. Mahalanobis distance analysis of the inflection points confirmed that changes in inflection point amplitude as a function of objects were greater than changes in timing. Inflection points were independent of the grasp force, consistent with the observation that reach-to-grasp kinematics and grasp force are controlled independently. In summary, the shaping of the hand during reach-to-grasp involves scaling the amplitude of highly stereotypic temporal movements of the fingers.