Video scene mutation change detection combined with SIFT algorithm
-
摘要: 视频场景变化检测对于视频的标注以及语义检索具有非常重要的作用。本文提出了一种结合SIFT(Scale Invariant Feature Transformation)特征点提取的场景变化检测算法。首先利用SIFT算法分别提取出视频前后帧的特征点并分别统计其数量,然后对视频前后帧进行图像匹配,统计匹配上的特征点数量,最后将该帧的匹配特征点数量与该帧前一帧的特征点数量做比值,从而通过该比值判断场景变化情况。实验结果表明,视频场景突变检测率平均可以达到95.79%。本算法可以在视频帧进行图像匹配的过程中对场景的变化情况进行判断,因此该算法不仅应用范围较广,还可以保证场景变化检测的精度,仿真结果证明了算法的有效性。Abstract: Video scene change detection has a very important role for video annotation and semantic search.This paper proposes a scene mutation change detection algorithm combined with SIFT(Scale Invariant Feature Transformation) feature point extraction.Firstly, the feature points of two adjacent video frames are extracted respectively using SIFT algorithm and the number of them is counted respectively.Then image matching of the two adjacent frames of the video is performed and the number of matching feature points is counted.Finally, the ratio between the number of matching feature points of the current frame and the number of matching feature points of its previous frame is calculated, so as to judge the scene change by this ratio.The average scene mutation change detection rate in the experimental results can reach 95.79%.The proposed algorithm can judge scene change during image matching.Therefore, the algorithm can not only be applied widely, but also guarantee the accuracy of scene change detection.Experimental results show the effectiveness of the proposed algorithm.
-
Key words:
- SIFT /
- feature point matching /
- scene change detection
-
图 2 无场景突变举例(115、116帧)
Figure 2. Examples of no scene mutation change(the 115th and 116th frames)
图 3 场景突变举例(114、115帧)
Figure 3. Examples of scene mutation change(the 114th and 115th frames)
表 1 实验数据举例
Table 1. Examples of experimental data
视频帧 特征点数量 与前一帧匹配的特征点数量 与前一帧特征点匹配率/% 是否发生场景突变 114 276 238 85.9 否 115 299 26 9.42 是 116 298 199 66.56 否 
下载: 导出CSV
表 2 SIFT特征点提取耗时统计
Table 2. Time consumption of SIFT feature point extraction
视频片段 总帧数 SIFT特征点提取所用时间/s 平均1 s完成特征点提取的视频帧数/s 动画(哆啦A梦) 1 922 72.717 6 26.431 0
下载: 导出CSV
表 3 视频中场景突变点
Table 3. Scene change points of videos in our experiments
视频片段 总帧数 场景突变点数目 电影 3 479 117 动画 1 922 100 新闻 2 180 39 MV 3 022 147
下载: 导出CSV
表 4 本算法实验结果
Table 4. Results obtained using the algorithm proposed in this paper
视频片段 Nc Nm Nf Pre/% Ppre/% 电影 112 5 10 95.73 91.80 动画 98 2 7 98.00 93.33 新闻 37 2 0 94.87 100.00 MV 138 8 11 94.56 92.67
下载: 导出CSV
-
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] -
下载:
计量
- 文章访问数: 1925
- HTML全文浏览量: 393
- PDF下载量: 1074
- 被引次数: 0
