Increasing interest, enthusiasm of sport lovers, and economics involved offer high importance to sports video recording and analysis. Being crucial for decision making, ball detection and tracking in soccer has become a challenging research area. This paper presents a novel deep learning approach for 2D ball detection and tracking (DLBT) in soccer videos posing various challenges. A new 2-stage buffer median filtering background modelling is used for moving objects blob detection. A deep learning approach for classification of an image patch into three classes, i.e. ball, player, and background is initially proposed. Probabilistic bounding box overlapping technique is proposed further for robust ball track validation. Novel full and boundary grid concepts resume tracking in ball_track_lost and ball_out_of_frame situations. DLBT does not require human intervention to identify ball from the initial frames unlike the most published algorithms. DLBT yields extraordinary accurate and robust tracking results compared to the other contemporary 2D trackers even in presence of various challenges including very small ball size and fast movements.

A deep learning ball tracking system in soccer videos

In recent years sport video research has gained a steady interest among the scientific community. The large amount of video data available from broadcast transmissions and from dedicated camera setups, and the need of extracting meaningful information from data, pose significant research challenges. Hence, computer vision and machine learning are essential for enabling automated or semi-automated processing of big data in sports. Although sports are diverse enough to present unique challenges on their own, most of them share the need to identify active entities such as ball or players. In this paper, an innovative deep learning approach to the identification of the ball in tennis context is presented. The work exploits the potential of a convolutional neural network classifier to decide whether a ball is being observed in a single frame, overcoming the typical issues that can occur dealing with classical approaches on long video sequences (e.g. illumination changes and flickering issues). Experiments on real data confirm the validity of the proposed approach that achieves 98.77% accuracy and suggest its implementation and integration at a larger scale in more complex vision systems.

/pdf/convolutional-neural-networks-based-ball-detection-in-tennis-m9pmrlosid.pdf

Convolutional Neural Networks Based Ball Detection in Tennis Games

Soccer broadcast video understanding has been drawing a lot of attention in recent years within data scientists and industrial companies. This is mainly due to the lucrative potential unlocked by effective deep learning techniques developed in the field of computer vision. In this work, we focus on the topic of camera calibration and on its current limitations for the scientific community. More precisely, we tackle the absence of a large-scale calibration dataset and of a public calibration network trained on such a dataset. Specifically, we distill a powerful commercial calibration tool in a recent neural network architecture on the large-scale SoccerNet dataset, composed of untrimmed broadcast videos of 500 soccer games. We further release our distilled network, and leverage it to provide 3 ways of representing the calibration results along with player localization. Finally, we exploit those representations within the current best architecture for the action spotting task of SoccerNetv2, and achieve new state-of-the-art performances.

/pdf/camera-calibration-and-player-localization-in-soccernet-v2-1j49d8vyv2.pdf

Camera Calibration and Player Localization in SoccerNet-v2 and Investigation of their Representations for Action Spotting

We present an automatic technique for calculating ball possession statistics from the video of a soccer match. The possession statistics is generated based on the number of valid passes made by an individual team. A valid pass is detected as a split or merge event of the ball with a player. A pass starts when the ball splits from a player. A pass ends when the ball merges with a player. We use a minimum-cost flow network to model number of valid passes in the soccer match. The ball and the players represent the nodes of the network. Each edge of the network is associated with a cost derived from the between-frame correspondences of the ball and the players. The total flow through the network is optimized to track the number of valid passes. Experimental results show that the accuracy of the proposed method is at least 4% better than that of a similar approach.

/pdf/generation-of-ball-possession-statistics-in-soccer-using-3a74b22ipe.pdf

Generation of Ball Possession Statistics in Soccer Using Minimum-Cost Flow Network

The paper describes a deep neural network-based detector dedicated for ball and players detection in high resolution, long shot, video recordings of soccer matches. The detector, dubbed FootAndBall, has an efficient fully convolutional architecture and can operate on input video stream with an arbitrary resolution. It produces ball confidence map encoding the position of the detected ball, player confidence map and player bounding boxes tensor encoding players' positions and bounding boxes. The network uses Feature Pyramid Network desing pattern, where lower level features with higher spatial resolution are combined with higher level features with bigger receptive field. This improves discriminability of small objects (the ball) as larger visual context around the object of interest is taken into account for the classification. Due to its specialized design, the network has two orders of magnitude less parameters than a generic deep neural network-based object detector, such as SSD or YOLO. This allows real-time processing of high resolution input video stream. Our code and pre-trained model can be found on the project website: this https URL .

FootAndBall: Integrated Player and Ball Detector.

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

Tracking of ball in sports videos is one of the most challenging tasks in computer vision and video processing domain. Recent ball tracking approaches fail to handle tracking of a small size and fast moving ball. Inaccurate 2D ball detection leads to further deterioration of 3D ball tracking results. This paper presents a soccer ball tracking by detection approach using a pre-trained Convolutional Neural Network (CNN). The proposed algorithm used CNN for identifying ball from background and other moving objects like players and referees. The 2D ball detection results are fine-tuned for identifying true ball positions. True ball positions, from cameras shooting the scene from different angle, are further mapped on ground plane. The actual ball movement is tracked in 3D from top-view. Experiments show that the proposed algorithm can tackle challenges like small ball size, shape changes, occlusion and tracking high-speed balls.

Paresh R. Kamble

Papers

Ball tracking in sports: a survey

A deep learning ball tracking system in soccer videos

A convolutional neural network based 3D ball tracking by detection in soccer videos