Fast Objectionable Videos Detection Algorithm Based on Motion Vector Field

摘要：

　　Aim at large amount and complex features of network video data,this paper proposed a fast algorithm of detecting pornographic video in streaming media.Combining the information of motion vector field and the features of pornographic behavior trajectory,we could judge that weather a video is pornographic or not without decoding it.The experiment show that the algorithm recognition rate is reached 80%,and is robust for network video detection environment.

关键词： Motion Vector Feature Extraction Reciprocating Motion Histogram H.264

作者: Lv Guiqing Qu Zhiyi

作者单位: School of Information Science & Engineering Lanzhou University line Lanzhou,China

会议类型: 国际会议

会议名称: 2013 2nd International Conference on Computer Science and Electronics Engineering(ICCSEE2013)(2013年第二届计算机科学与电子工程国际会议)

会议地点: 杭州

会议语种:英文

页码: 1970-1973

在线出版日期: 2013-03-22（万方平台首次上网日期，不代表论文的发表时间）