Fig 10 - uploaded by Yongduek Seo
Content may be subject to copyright.
Source publication
This paper deals with tracking players and the ball in multiple soccer video sequences taken from fixed cameras located around the stadium. One of the major obstacle in player tracking occurs when occlusion between people results in insufficient observation information. If a player blob is not visible in some of the input videos due to occlusion, o...
Similar publications
We describe a novel scalable approach for the management of a large number of Pan-Tilt-Zoom (PTZ) cameras deployed outdoors for persistent tracking of humans and vehicles, without resorting to the large fields of view of associated static cameras. Our system, Active Collaborative Tracking - Vision (ACT-Vision), is essentially a real-time operating...
Motion segmentation (MS) is an essential step in video analysis. Its quantitative and qualitative evaluation is largely dependent on the dataset used for testing. Although there are publicly available datasets such as Hopkins and FBMS, they have limitations in terms of number of motions, partial/complete occlusion, stopping motion, sequence length,...
Object tracking is an important research direction in computer vision and is widely used in video surveillance, security monitoring, video analysis and other fields. Conventional tracking algorithms perform poorly in specific scenes, such as a target with fast motion and occlusion. The candidate samples may lose the true target due to its fast moti...
Robot person following (RPF) is a capability that supports many useful human-robot-interaction (HRI) applications of a mobile robot. However, existing solutions to person following often assume a full observation of the tracked person. As a consequence, they cannot track the person reliably under partial occlusion where the assumption of full obser...
This paper deals with a problem of tracking players in multiple soccer video sequences taken from fixed cameras located around the stadium. One of the major obstacle in player tracking occurs when occlusion between people results in not enough observation information. If a player blob is not visible in some of the input videos due to occlusion, one...
Citations
... This work used homography projection to form a background template over all frames by computing a Gaussian modal of each pixel. [17] used homography transformation between multiple observed frames to locate ball and players. [18] presented a real-time computer vision system that tracks the motion of a tennis ball in 3D using multiple cameras. ...
In this paper we describe a framework to improve the detection of ball hit events in tennis games by combining audio and visual information. Detection of the presence and timing of these events is crucial for the understanding of the game. However, neither modality on its own gives satisfactory results: audio information is often corrupted by noise and also suffers from acoustic mismatch between the training and test data, and visual information is corrupted by complex backgrounds, camera calibration, and the presence of multiple moving objects. Our approach is to first attempt to track the ball visually and hence estimate a sequence of candidate positions for the ball, and to then locate putative ball hits by analysing the ball's position in this trajectory. To handle the severe interferences caused by false ball candidates, we smooth the trajectory by using locally weighted linear regression and removing the frames where there are no candidates. We use Gaussian mixture models to generate estimates of the times of hits using the audio information, and then integrate these two sources of information in a probabilistic framework. Testing our approach on three complete tennis games shows significant improvements in detection over a range of conditions when compared with using a single modality.
Monitoring of player activity within a competition is currently a reality within some high performance sporting teams, and the demand and level of sophistication for this information will continue to grow as players and coaches seek this knowledge. This new found scientific wisdom can guide the training and preparation of the athlete, with the aim of improving performance and reducing the likelihood of injuries. To date this information has been collected manually, which is time consuming and expensive. The challenges are to validate the accuracy of these systems and once this criterion is satisfied to expedite the analysis process to enable as close to real time feedback as possible. Outside of the coach and player arena there is growing demand from the other associated parties, such as the host broadcast media, referees and the spectator, all of whom are seeking new knowledge on "what is happening" during the play.
In a sports video, the significant events are caused mostly because of ball-player and player-player interaction. To detect and track a ball or a player in a sports video becomes more challenging in presence of many moving objects in the background. The small size of the ball in relation to the frame size makes ball detection much more difficult. In addition, the ball-images are also getting deformed due to the high speed movement of the ball. Often the ball gets occluded by players and the ball image gets merged with lines and boarders in the field. In this paper, the problem of ball detection-and-tracking in a real time basketball video is addressed. Here a trajectory based ball-detection-and-tracking method is proposed to detect and track the ball in a set of videos of basketball long shot sequences. A two-fold detection-and-tracking framework is used where the ball is detected using a feature-based method in the first step. The second step verifies the detection result of the ball detection system using the 2D-trajectory information of the possible ball candidates in the frame. The true ball trajectory is extracted from a set of candidate trajectories and the ball locations are estimated along the trajectory. The missing ball positions due to occlusion of the ball and merging of the ball image with other objects in the background are estimated using a trajectory interpolation technique.
