1. SCSIT
  2. All Volumes and Issues
  3. Volume 751, 2013
  4. Key frame recognition using voronoi tessellations

Key frame recognition using voronoi tessellations

SCSIT.
2013;
Volume 751
: pp. 52 - 58
Authors: 

S. Mashtalir, O. Mikhnova

Kharkiv National University of Radio Electronics, Informatics department

Вилучення ключових кадрів є формою скорочення відеоматеріалу. Цю задачу пропонується вирішувати за допомогою діаграм Вороного, які будуються за опорними точками. Вибирати ці точки пропонується за методом Харріса. Для того, щоб покращити розміщення точок, використовувалась кластеризація методом k-середніх. У результаті опорні точки значно більше відповідають контенту відео, що допомагає, своєю чергою, вилучати лише значущі кадри.

Key frame extraction is a form of video summarization. It is proposed to be performed with Voronoi diagrams which are constructed on salient points. Salient point selection is assumed to be done via Harris method. To perfect point location, k-means clustering is used. As a result, salient points much better correspond to video content, which helps to extract meaningful frames.

salient point
key frame
video summarization
Voronoi diagram
k-means clustering
Harris method
  1. Szeliski R. Computer Vision. Algorithms and Applications. – London: Springer, 2011. – 813 p.
  2. Goldman D.R. Framework for video annotation, visualization, interaction: Doctoral Thesis … Doctor of Philosophy. – Washington, 2007. – 140 p.
  3. Zhang D., Liu Y., Hou J. Digital Image Retrieval Using Intermediate Semantic Features and Multistep Search. In: Digital Image Computing: Techniques and Applications. – 2008. – pp. 513-518.
  4. Jiang Y.-G., Ngo C.-W., Yang J. Towards Optimal Bag-of-Features for Object Categorization and Semantic Video Retrieval. In: 6th ACM International Conference on Image and Video Retrieval. – 2007. – pp. 494–501.
  5. Tsoneva Т. Automated summarization of movies and TV series on a semantic level: Doctoral Thesis … Doctor of Philosophy. – Eindhoven, 2007. – 126 p.
  6.  Bodyanskiy Y. et al. On-line video segmentation using methods of fault detection in multidimensional time sequences. In: International Journal of Electronic Commerce Studies. – 2012. – Vol. 3, No. 1. – pp. 1–20.
  7. Mikhnova O. A template-based approach to key frame extraction from video. In: International scientific and technical Internet conf. Computer Graphics and Image Recognition. – Vinnytsia: VNTU. – 2012. – pp. 120–127.
  8. Sebe N., Lew M.S. Comparing salient point detectors. In: Pattern Recognition Letters. – 2003. – Vol. 24, No. 1–3. – pp. 89–96.
  9. Tsai Y.-H. Salient points reduction for content-based image retrieval. In: World Academy of Science, Engineering and Technology. – 2009. – Vol. 49. – pp. 656–659.
  10. Du Q., Faber V., Gunzburger M. Centroidal Voronoi tessellations: Applications and algorithms. In: Society for Industrial and Applied Mathematics Review. – 1999. – Vol. 41, No. 4. – pp. 637–676.
  11. Hurtado F. et al. The weighted farthest color Voronoi diagram on trees and graphs. In: Computational Geometry: Theory and Applications. – 2004. – Vol. 27, No. 1. – pp. 13-26.
  12. Mashtalir V. et al. A novel metric on partitions for image segmentation. In: IEEE International Conference on Video and Signal Based Surveillance. – 2006 – 6 p.
  13. Gonzalez R., Woods R., Eddins S. Digital image processing using MatLab. – Upper Sadle River, NJ: Pearson Prentice Hall, 2004. – 609 p.
  14. Sonka M., Hlavac V., Boyle R. Image processing, analysis, and machine vision, International student edition. – 3rd ed. – Toronto: Thomson, 2007. – 850 p.
  15. Bezdek J.C. et al. Fuzzy models and algorithms for pattern recognition and image processing. – New York: Springer, 2005. – 776 p.
  16. Schonfeld D. et al. (Eds.) Video search and mining. In: Studies in Computational Intelligence. – Berlin: Springer. – 2010. – Vol. 287. – 388 p.
  17. Haralick R.M., Shanmugam K., Dinstein I. Textural features for image classification. In: IEEE Transactions on Systems, Man and Cybernetics. – 1973. – Vol. 3, No. 6. – pp. 610–621.