收录:
摘要:
Motions of a camera can reflect certain kinds of semantic information in video sequences; therefore, in this paper, according to the distant court shot in basketball game, we calculate the motion vector first, then analyze the semantic information of the motion vector, and finally annotate the offense-defense actions in games. Experiments are showing the possibility that the proposed technique can annotate the offensedefense semantic actions as expected. ©2009 IEEE.
关键词:
通讯作者信息:
电子邮件地址:
来源 :
年份: 2009
页码: 146-149
语种: 英文