There is a large variety of trackers, which have been proposed in the literature during the last two decades with some mixed success. Object tracking in realistic scenarios is a difficult problem, therefore, it remains a most active area of research in computer vision. A good tracker should perform well in a large number of videos involving illumination changes, occlusion, clutter, camera motion, low contrast, specularities, and at least six more aspects. However, the performance of proposed trackers have been evaluated typically on less than ten videos, or on the special purpose datasets. In this paper, we aim to evaluate trackers systematically and experimentally on 315 video fragments covering above aspects. We selected a set of nineteen trackers to include a wide variety of algorithms often cited in literature, supplemented with trackers appearing in 2010 and 2011 for which the code was publicly available. We demonstrate that trackers can be evaluated objectively by survival curves, Kaplan Meier statistics, and Grubs testing. We find that in the evaluation practice the F-score is as effective as the object tracking accuracy (OTA) score. The analysis under a large variety of circumstances provides objective insight into the strengths and weaknesses of trackers.

/pdf/visual-tracking-an-experimental-survey-50w2t40998.pdf

Visual Tracking: An Experimental Survey

Sports data analysis is becoming increasingly large scale, diversified, and shared, but difficulty persists in rapidly accessing the most crucial information. Previous surveys have focused on the methodologies of sports video analysis from the spatiotemporal viewpoint instead of a content-based viewpoint, and few of these studies have considered semantics. This paper develops a deeper interpretation of content-aware sports video analysis by examining the insight offered by research into the structure of content under different scenarios. On the basis of this insight, we provide an overview of the themes particularly relevant to the research on content-aware systems for broadcast sports. Specifically, we focus on the video content analysis techniques applied in sportscasts over the past decade from the perspectives of fundamentals and general review, a content hierarchical model, and trends and challenges. Content-aware analysis methods are discussed with respect to object-, event-, and context-oriented groups. In each group, the gap between sensation and content excitement must be bridged using proper strategies. In this regard, a content-aware approach is required to determine user demands. Finally, this paper summarizes the future trends and challenges for sports video analysis. We believe that our findings can advance the field of research on content-aware video analysis for broadcast sports.

A Survey of Content-Aware Video Analysis for Sports

We employ hierarchical data association to track players in team sports. Player movements are often complex and highly correlated with both nearby and distant players. A single model would require many degrees of freedom to represent the full motion diversity and could be difficult to use in practice. Instead, we introduce a set of Game Context Features extracted from noisy detections to describe the current state of the match, such as how the players are spatially distributed. Our assumption is that players react to the current situation in only a finite number of ways. As a result, we are able to select an appropriate simplified affinity model for each player and time instant using a random decision forest based on current track and game context features. Our context-conditioned motion models implicitly incorporate complex inter-object correlations while remaining tractable. We demonstrate significant performance improvements over existing multi-target tracking algorithms on basketball and field hockey sequences several minutes in duration and containing 10 and 20 players respectively.

/pdf/tracking-sports-players-with-context-conditioned-motion-2hvql9mr00.pdf

Tracking Sports Players with Context-Conditioned Motion Models

Increase in the number of sport lovers in games like football, cricket, etc. has created a need for digging, analyzing and presenting more and more multidimensional information to them. Different classes of people require different kinds of information and this expands the space and scale of the required information. Tracking of ball movement is of utmost importance for extracting any information from the ball based sports video sequences. Based on the literature survey, we have initially proposed a block diagram depicting different steps and flow of a general tracking process. The paper further follows the same flow throughout. Detection is the first step of tracking. Dynamic and unpredictable nature of ball appearance, movement and continuously changing background make the detection and tracking processes challenging. Due to these challenges, many researchers have been attracted to this problem and have produced good results under specific conditions. However, generalization of the published work and algorithms to different sports is a distant dream. This paper is an effort to present an exhaustive survey of all the published research works on ball tracking in a categorical manner. The work also reviews the used techniques, their performance, advantages, limitations and their suitability for a particular sport. Finally, we present discussions on the published work so far and our views and opinions followed by a modified block diagram of the tracking process. The paper concludes with the final observations and suggestions on scope of future work.

Ball tracking in sports: a survey

Robotic dance is an important topic in the field of social robotics. Its research has a vital significance to both humans and robotics. This paper presents a review of the state of the art in robotic dance. Robotic dance is classified into four categories: cooperative human–robot dance, imitation of human dance motions, synchronization for music, and creation of robotic choreography. The research methods in each category are discussed. Future research areas are highlighted.

Robotic Dance in Social Robotics—A Taxonomy

Recent years have brought an increasing interest on analyzing efficiently the performance of sports players during training sessions and games. The information collected from such analysis is very valuable to educators and coaches since it allows them to better understand the difficulties of a trainee, a player or even an entire team and formulate adequate training and strategic plans accordingly. In order to perform this analysis in a consistent and systematic way, sophisticated sensory systems and data processing techniques are needed. This paper presents a survey on relevant work, current techniques and trends on the area of team tracking systems applied to sports. We propose a classification of these systems by distinguishing them into two main categories: intrusive and nonintrusive. Nonintrusive systems are further refined into outdoor and indoor sports applications. The specific characteristics of each system are itemized, including the identification of the strong points and limitations. Finally, the paper highlights some open issues and research opportunities on this area.

Survey on team tracking techniques applied to sports

Assessing physical activity (PA) is a challenging task and many different approaches have been proposed. Direct observation (DO) techniques can objectively code both the behavior and the context in which it occurred, however, they have significant limitations such as the cost and burden associated with collecting and processing data. Therefore, this study evaluated the utility of an automated video analysis system (CAM) designed to record and discriminate the intensity of PA using a subject tracking methodology. The relative utility of the CAM system and DO were compared with criterion data from an objective accelerometry-based device (Actigraph GT3X+). Eight 10 year old children (three girls and five boys) wore the GT3X+ during a standard basketball session. PA was analyzed by two observers using the SOPLAY instrument and by the CAM system. The GT3X+ and the CAM were both set up to collect data at 30 Hz while the DO was performed every two minutes, with 10 s of observation for each gender. The GT3X+ was processed using cut points by Evanson and the outcome measure was the percentage of time spent in different intensities of PA. The CAM data were processed similarly using the same speed thresholds as were used in establishing the Evenson cut-off points (light: 4 mph). Similar outcomes were computed from the SOPLAY default analyses. A chi-square test was used to test differences in the percentage of time at the three intensity zones (light, walking and very active). The Yates' correction was used to prevent overestimation of statistical significance for small data. When compared with GT3X+, the CAM had better results than the SOPLAY. The chi-square test yielded the following pairwise comparisons: CAM versus GT3x+ was χ2 (5) = 24.18, p < .001; SOPLAY2 versus GT3x+ was χ2 (5) = 144.44, p < .001; SOPLAY1 versus GT3x+ was χ2 (5) = 119.55, p < .001. The differences were smaller between CAM and GT3x+, suggesting that the video tracking system provided better agreement than DO. The small sample size precludes a definitive evaluation but the results show that the CAM video system may have promise for automated coding of physical activity behavior.

/pdf/assessing-physical-activity-intensity-by-video-analysis-2npktjgl0m.pdf

Assessing physical activity intensity by video analysis.

In recent years there has been a growing interest by the sport’s experts (teachers and coaches) in developing automatic systems for detecting, tracking and identifying player’s movements with the purpose of improving the players’ performance and accomplishing a consistent and standard analysis of the game metrics. A challenge like this requires sophisticated techniques from the areas of image processing and artificial intelligence. The objective of our work is to study and develop hardware and software techniques in order to build an automatic visual system for detecting and tracking players in indoor sports games that can aid coaches to analyse and improve the players’ performance. Our methodology is based on colour features and therefore several colour image processing techniques such as background subtraction, blob colour definition (RGB and HSL colour spaces) and colour blob manipulation are employed in order to detect the players. Past information and players’ velocity allow the tracking algorithm to define probable areas. Tests conducted with a single IP surveillance camera on the sports hall of the Faculty of Sports from the University of Porto showed detection rates from 72.2% to 93.3%.

/pdf/real-time-colour-based-player-tracking-in-indoor-sports-4a5aq3v2z3.pdf

Real Time Colour Based Player Tracking in Indoor Sports

This paper presents a system that is able to control in real time a humanoid robot to perform dance movements to the rhythm of a music. The movements' coordination is performed with the aid of a music analyzer that estimates the beat of the music and calculates a prediction of the next inter-beat-interval (IBI). During the robot dancing performance, a dance movement is chosen from a pre-defined dance library and is executed on-the-fly by the robot. The movements' velocity, as well as the attended metrical-level, are adjusted in real time so that the movement can be executed within the time interval of two beats, and at the same time taking into consideration the robot motor-rates limitations. Results evince a good synchrony of the movement towards the analyzed rhythm, but points to the need of a careful design of the dance movements so they can be naturally executed on time.

Autonomous robot dancing synchronized to musical rhythmic stimuli

The sports community needs technological aid to extract accurate statistics and performance data from both practice sessions and games. To obtain such information, players must be tracked over time and their movements processed so that individual actions and team plays are simultaneously analyzed. In order to perform this analysis in an automated, formal and accurate way, the authors developed a cost conscientious processing system fed by two overhead cameras (roughly one video stream for each half-field). Players are detected by vest colors, and Fuzzy Logic is used to allow for a given color to be shared by different teams. Color models for the background and the teams are dynamic over time to make up for changes in natural lighting conditions and consequent color changes. Player tracking is further enhanced using Kalman Filtering. Some examples of the analysis, made possible by the proposed system, are shown. Results are based on videos collected during the Portuguese Handball SuperCup competition for the year 2011.

/pdf/vision-system-for-tracking-handball-players-using-fuzzy-260dajx3wi.pdf

Catarina B. Santiago

Papers

Survey on team tracking techniques applied to sports

Assessing physical activity intensity by video analysis.

Real Time Colour Based Player Tracking in Indoor Sports

Autonomous robot dancing synchronized to musical rhythmic stimuli

Vision system for tracking handball players using fuzzy color processing