educative.io

How does the video duplication find what video to compare against?

Courses
,
#1

“As soon as a user starts uploading a video, our service can run video matching algorithms (e.g. Block Matching) to find duplicates”

@Design_Gurus
Question : How exactly will we know what video to compare the current video being uploaded against ? (What video will we match the current video against to run duplication detection? )

#2

Hi @Rohan_Joshi
YouTube implemented a solution called “Content ID” that is based on criteria for comparison including

  • the size of the video
  • the format the video was uploaded in
  • frame rate
  • audio waveform
  • type of encoding
  • pixel distribution etc.
    Each uploaded video gets a content ID based on these criteria and can then be used to determine if there is an audio or video match or even a partial match. Content ID is a digital “fingerprint” of a work of art (video + audio) that is practical in size and allows fast comparisons. These Content IDs are stored in a database. When a new video is uploaded to YouTube, YouTube compares the Content ID of the newly uploaded video with the values already in the database. If there’s a good enough match, the newly added video is flagged for additional analysis/comparison.
    Hope it will help, Happy Learning :slight_smile: