Scorebox Extraction from Mobile Sports Videos usingSupport Vector Machines

摘要

Scorebox plays an important role in understanding contents of sports videos. However, the tiny scorebox may givethe small-display-viewers uncomfortable experience in grasping the game situation. In this paper, we propose a novelframework to extract the scorebox from sports video frames. We first extract candidates by using accumulated intensityand edge information after short learning period. Since there are various types of scoreboxes inserted in sports videos,multiple attributes need to be used for efficient extraction. Based on those attributes, the optimal information gain iscomputed and top three ranked attributes in terms of information gain are selected as a three-dimensional feature vectorfor Support Vector Machines (SVM) to distinguish the scorebox from other candidates, such as logos and advertisementboards. The proposed method is tested on various videos of sports games and experimental results show the efficiencyand robustness of our proposed method.