Improving the robustness of motion vector temporal descriptor

Farzaneh Rahmani, Farzad Zargari, Mohammad Ghanbari
  • IET Image Processing, January 2018, the Institution of Engineering and Technology (the IET)
  • DOI: 10.1049/iet-ipr.2017.0206

Making motion vector descriptor robust against variations in resolutions and aspect ratios of videos

What is it about?

Motion vector descriptors are used to find the similarity of videos, but they fail to demonstrate the similarity of videos with different resolutions or aspect ratios. In this paper the required process to improve the robustness of motion vector against resolution and aspect ratio is presented. The process is computationally simple and experimental results indicate considerable improvement by the proposed method in finding similar videos and discriminating various videos with different resolution and aspect ratio, compared with original motion vector descriptors.

Why is it important?

By using the proposed method the modified motion vectors can effectively determine similar videos and reliably discriminate videos with different aspect ratios and resolutions.

Perspectives

Farzad Zargari (Author)
Research Institute for ICT (aka Iran Telecom Recearch Center)

In a project about automatic evaluation of video search engines we should find the similarity of various videos with different resolution and aspect ratios and we considered that the original motion vector descriptor does not perform well in these cases. This paper is the result of our efforts to remove this weakness.

Read Publication

http://dx.doi.org/10.1049/iet-ipr.2017.0206

The following have contributed to this page: Farzad Zargari