Key Frame Extraction Based on Motion Vector
Video key frame extraction is a video summarization technique which is used to recapitulate and describe the most important segments of video sequence. Video clips containing motion information are more likely to draw users’ attention. Accordingly, we propose a novel key frame extraction scheme based on motion vector. A video sequence is partitioned into shots by distance between consecutive video frames calculated by Relative Entropy (RE). Difference of magnitude of motion vector between neighboring video images within a shot is computed to localize video clips containing significant content changes. And such segments are defined as active sub-shots in this paper. The key frames are extracted from each active sub-shots and other inactive sub-shots by exploiting different algorithms. Experimental results show that our proposed method obtains exact and complete results in video key frame extraction.
KeywordsKey frame extraction Motion vector
This work has been funded by Natural Science Foundation of China (61471261, 61179067, U1333110), and by grants TIN2013-47276-C6-1-R from Spanish Government and 2014-SGR-1232 from Catalan Government (Spain).
- 9.Ma, Y.F., Lu, L., Zhang, H.J., Li, M.: A user attention model for video summarization. In: Proceedings of the Tenth ACM International Conference on Multimedia, pp. 533–542. ACM (2002)Google Scholar
- 10.Mentzelopoulos, M., Psarrou, A.: Key-frame extraction algorithm using entropy difference. In: Proceedings of the 6th ACM SIGMM International Workshop on Multimedia Information Retrieval, pp. 39–45. ACM. New York, October 2004Google Scholar