A Fingerprint Generation based Method for Video Authenticity

Authors

  • Jaspreet Kaur
  • Avinash Sharma

Abstract

The expanding interest for copyright assurance has gotten broad intrigue video fingerprinting innovation, which is considered as a viable path for computerized content confirmation. In this paper, a creative strategy for video credibility dependent on the unique mark of scene casing has been proposed. Since the likelihood of 5 indistinguishable scene outlines existing progressively in various recordings is incredibly low, the proposed strategy right off the bat separate 5 progressive distinctive scene edges to make up the video unique finger impression. This video unique mark is at that point joined with the ID data of video to shape meta unique finger impression information. The metadata is put away in Bag-words group, prompting the sparing of 75% extra room. During the time spent validation, a double search of modified document is connected to quicken the coordinating velocity. The consequences of reenactment analyses utilizing MATLAB demonstrate that the video validness technique proposed in this paper accomplishes brilliant location execution, with the normal precision over 98% at a normal speed of 12 second for each video. This causes the constant location to turn into conceivable on the Internet.

Downloads

Published

2020-03-14

Issue

Section

Articles