摘要
视频敏感字幕信息的出现严重影响着电视节目安全播出。针对此问题,本文提出了一种基于近邻传播聚类的视频字幕监测方法。首先分割完整视频至各个独立画面;然后对其分别累积边缘图像,并基于近邻传播聚类算法进行字幕边缘划分以得到边缘子图;进而采用二值化,去燥获得清晰的字幕子图;最后OCR工具完成字幕识别。此外,为了验证本方法的有效,三组带字幕视频节目被用于实验,实验结果表明三组视频的敏感字幕检测率均达到87%以上。
The appearance of video sensitive information seriously affects the safe broadcasting of TV programs.To solve this problem,a video subtitle monitoring method based on affinity propagation clustering is proposed in this paper.First,the complete video is divided into each independent image;then the edge image is accumulated respectively,and the edge sub image is divided based on the affinity propagation clustering algorithm.Then the binaryzation and remove noises are used to get a clear sub image.Finally,the OCR tool completes the subtitle recognition.In addition,in order to verify the effectiveness of this method,three groups of video with subtitle are used for experiments.The experimental results show that the detection rate of sensitive subtitles for three group's video is up to 87%.
引文
[1]金阳,程江华,任通,等.基于视频帧加权的台标自动检测方法[J].电视技术,2012,36(21):33-36.
[2]赵亚男,王履程,张冰涛.基于视频分析的电视视频监测技术[J].广播与电视技术,2017,44(10):128-131.
[3]张立山,倪云飞.浅谈广播电视技术监测现状与发展[J].科技致富向导,2012(29):423-423.
[4]庄越挺,刘骏伟,吴飞,等.基于支持向量机的视频字幕自动定位与提取[J].计算机辅助设计与图形学学报,2002,14(8):750-753.
[5]谌先敢,刘娟,高智勇,等.基于累积边缘图像的现实人体动作识别[J].自动化学报,2012,38(8):1380-1384.
[6]易剑,彭宇新,肖建国.基于颜色聚类和多帧融合的视频文字识别方法[J].软件学报,2011,22(12):2919-2933.
[7]SERDAH A M,ASHOUR W M.Clustering Large-Scale Data Based On Modified Affinity Propagation Algorithm[J].Journal of Artificial Intelligence&Soft Computing Research,2016,6(1):23-33.
[8]张冰涛,王小鹏,王履程,等.基于图论的MANET入侵检测方法[J].电子与信息学报,2018,45(6):1086-1093.
[9]GATOS B,PRATIKAKIS I,PERANTONIS S J.Adaptive degraded document image binarization[J].Pattern Recognition,2006,39(3):317-327.
[10]WOLF C,JOLION J M.Extraction and recognition of artificial text in multimedia documents[J].Formal Pattern A-nalysis and Applications,2004,6(4):309-326.
[11]OCR文字识别软件[CP/OL].http://xiazai.zol.com.cn/heji/6910/.