Motion analysis for broadcast tennis video considering mutual interaction of players
Naoto Maruyama · 2011
In this paper, we propose a new scheme of player recognition based on Cubic High-order Local Auto-Correlation (CHLAC) features. To achieve a high classification rate based on CHLAC features, various types of CHLAC features should be used, which are generated by controlling their parameters. However, some CHLAC features are unreliable for classification. To find the best CHLAC features, we apply the AdaBoost algorithm. Further, we add the information on the opposing player for classification enhancement. There are strong interactions between the actions of two competing tennis players, and thus it is effective to utilize their correlation for classification. Our approach of considering two player’s interactions achieved a classification success rate of 96.09%, which is much more accurate than methods using only the target player’s information, with which the classification success rate is 85.02%. 1