This paper analyzes the network of passes among the players of the Spanish team during the last FIFA World Cup 2010,where they emerged as the champion,with the objective of explaining the results obtained from the beh...This paper analyzes the network of passes among the players of the Spanish team during the last FIFA World Cup 2010,where they emerged as the champion,with the objective of explaining the results obtained from the behavior at the complex network level.The team is considered a network with players as nodes and passes as(directed) edges.A temporal analysis of the resulting passes network is also done,looking at the number of passes,length of the chain of passes,and to network measures such as player centrality and clustering coefficient.Results of the last three matches(the decisive ones) indicate that the clustering coefficient of the pass network remains high,indicating the elaborate style of the Spanish team.The effectiveness of the opposing team in negating the Spanish game is reflected in the change of several network measures over time,most importantly in drops of the clustering coefficient and passing length/speed,as well as in their being able in removing the most talented players from the central positions of the network.Spain's ability to restore their combinative game and move the focus of the game to offensive positions and talented players is shown to tilt the balance in favor of the Spanish team.展开更多
Football is one of the most-watched sports,but analyzing players’per-formance is currently difficult and labor intensive.Performance analysis is done manually,which means that someone must watch video recordings and ...Football is one of the most-watched sports,but analyzing players’per-formance is currently difficult and labor intensive.Performance analysis is done manually,which means that someone must watch video recordings and then log each player’s performance.This includes the number of passes and shots taken by each player,the location of the action,and whether or not the play had a successful outcome.Due to the time-consuming nature of manual analyses,interest in automatic analysis tools is high despite the many interdependent phases involved,such as pitch segmentation,player and ball detection,assigning players to their teams,identifying individual players,activity recognition,etc.This paper proposes a system for developing an automatic video analysis tool for sports.The proposed system is the first to integrate multiple phases,such as segmenting the field,detecting the players and the ball,assigning players to their teams,and iden-tifying players’jersey numbers.In team assignment,this research employed unsu-pervised learning based on convolutional autoencoders(CAEs)to learn discriminative latent representations and minimize the latent embedding distance between the players on the same team while simultaneously maximizing the dis-tance between those on opposing teams.This paper also created a highly accurate approach for the real-time detection of the ball.Furthermore,it also addressed the lack of jersey number datasets by creating a new dataset with more than 6,500 images for numbers ranging from 0 to 99.Since achieving a high perfor-mance in deep learning requires a large training set,and the collected dataset was not enough,this research utilized transfer learning(TL)to first pretrain the jersey number detection model on another large dataset and then fine-tune it on the target dataset to increase the accuracy.To test the proposed system,this paper presents a comprehensive evaluation of its individual stages as well as of the sys-tem as a whole.展开更多
基金supported in part by the CEI BioTIC GENIL(CEB09-0010)MICINN CEI Program(PYR2010-13)projectthe Andalusian Regional Government P08-TIC-03903,P08-TIC-03928,and TIC-6083 projectsMICINN projects TIN2008-05941 and TIN2011-28627-C04
文摘This paper analyzes the network of passes among the players of the Spanish team during the last FIFA World Cup 2010,where they emerged as the champion,with the objective of explaining the results obtained from the behavior at the complex network level.The team is considered a network with players as nodes and passes as(directed) edges.A temporal analysis of the resulting passes network is also done,looking at the number of passes,length of the chain of passes,and to network measures such as player centrality and clustering coefficient.Results of the last three matches(the decisive ones) indicate that the clustering coefficient of the pass network remains high,indicating the elaborate style of the Spanish team.The effectiveness of the opposing team in negating the Spanish game is reflected in the change of several network measures over time,most importantly in drops of the clustering coefficient and passing length/speed,as well as in their being able in removing the most talented players from the central positions of the network.Spain's ability to restore their combinative game and move the focus of the game to offensive positions and talented players is shown to tilt the balance in favor of the Spanish team.
文摘Football is one of the most-watched sports,but analyzing players’per-formance is currently difficult and labor intensive.Performance analysis is done manually,which means that someone must watch video recordings and then log each player’s performance.This includes the number of passes and shots taken by each player,the location of the action,and whether or not the play had a successful outcome.Due to the time-consuming nature of manual analyses,interest in automatic analysis tools is high despite the many interdependent phases involved,such as pitch segmentation,player and ball detection,assigning players to their teams,identifying individual players,activity recognition,etc.This paper proposes a system for developing an automatic video analysis tool for sports.The proposed system is the first to integrate multiple phases,such as segmenting the field,detecting the players and the ball,assigning players to their teams,and iden-tifying players’jersey numbers.In team assignment,this research employed unsu-pervised learning based on convolutional autoencoders(CAEs)to learn discriminative latent representations and minimize the latent embedding distance between the players on the same team while simultaneously maximizing the dis-tance between those on opposing teams.This paper also created a highly accurate approach for the real-time detection of the ball.Furthermore,it also addressed the lack of jersey number datasets by creating a new dataset with more than 6,500 images for numbers ranging from 0 to 99.Since achieving a high perfor-mance in deep learning requires a large training set,and the collected dataset was not enough,this research utilized transfer learning(TL)to first pretrain the jersey number detection model on another large dataset and then fine-tune it on the target dataset to increase the accuracy.To test the proposed system,this paper presents a comprehensive evaluation of its individual stages as well as of the sys-tem as a whole.
